MirBot, a collaborative object recognition system for smartphones using convolutional neural networks
MirBot is a collaborative application for smartphones that allows users to perform object recognition. This app can be used to take a photograph of an object, select the region of interest and obtain the most likely class (dog, chair, etc.) by means of similarity search using features extracted from a convolutional neural network (CNN). The answers provided by the system can be validated by the user so as to improve the results for future queries. All the images are stored together with a series of metadata, thus enabling a multimodal incremental dataset labeled with synset identifiers from the WordNet ontology. This dataset grows continuously thanks to the users’ feedback, and is publicly available for research. This work details the MirBot object recognition system, analyzes the statistics gathered after more than four years of usage, describes the image classification methodology, and performs an exhaustive evaluation using handcrafted features, convolutional neural codes and different transfer learning techniques. After comparing various models and transformation methods, the results show that the CNN features maintain the accuracy of MirBot constant over time, despite the increasing number of new classes. The app is freely available at the Apple and Google Play stores.
keywords:Object recognition, image retrieval databases, convolutional neural networks, deep learning, human computer interaction
Object recognition is a highly active topic in computer vision and can be particularly useful for mobile devices (Bock2010IImage:IPhone, ; mobilenets, ) as regards retrieving information about objects on the fly. Visually impaired persons can also benefit from these systems (Matusiak2013ObjectUsers, ).
MirBot is a smartphone app that users can train to recognize any object. The objects are categorized according to lemmas (such as chair, dog, laptop, etc.) selected from the WordNet lexical database (Fellbaum1998WordNet:Database, ). A user can employ MirBot to take a photograph and select a rectangular region of interest (ROI) in which the target object is located. The image, the ROI coordinates and a series of associated metadata are then sent to a server, which performs a similarity search using k-Nearest Neighbors (kNN) and returns the class of the most likely image, as shown in Figure 1. The app users can validate the system response in order to improve the classification results for future queries, and this feedback allows the database to grow continuously with new labeled images.
MirBot is designed as a pedagogical game in which a simulated robot can be trained in order to involve users in an automatic learning process. As pointed out in Barrington2012Game-poweredLearning. (); vonAhn2004LabelingGame (), developing machine learning tasks through games has proven to be a successful approach to make users enjoy labeling data. From the users’ perspective, the main distinctive feature of MirBot with regard to other apps is that it allows them to train a personal image search system, thus making the dataset dynamic and user driven.
This work extends the contents of the MirBot system for object retrieval introduced in MirBot:System (), which initially used handcrafted visual descriptors. The main contributions of this paper with regard to the initial work are a detailed description of the user interaction process, the statistics related to the database gathered after four years of usage, a new classification methodology based on CNN features (neural codes) obtained from pre-trained and fine-tuned models, and the evaluation results and conclusions.
When a user submits a photograph, visual descriptors are extracted and compared to the existing prototypes in the dataset in order to predict the class of the object. In the initial MirBot version MirBot:System (), both local features and color histograms were extracted and combined to obtain the most likely class. In this work, several convolutional neural network (CNN) features have been added to these descriptors for use in evaluation. The gap between the results obtained using handcrafted descriptors and features extracted from convolutional networks led the traditional image descriptors in MirBot to be replaced with neural codes in June 2015.
As pointed out in Torralba200880Recognition. (), finding images within large collections is an important topic for the computer vision community. Recent progress in object recognition has been built upon efforts to create large-scale, real-world image datasets (Khosla2012UndoingBias, ; Deng2009ImageNet:Database, ) that are crucial for developing robust image retrieval algorithms, in addition to considering the large amount of data required in recent deep neural networks (LeCun2015DeepLearning, ).
One of the main contributions of MirBot is a dataset with a similar structure to that of ImageNet (Deng2009ImageNet:Database, ), with the exception that all the images are gathered with smartphone cameras and stored with their associated metadata and with regions of interest. In October 2016 we had validated images distributed in classes. Although the MirBot dataset still cannot be considered a very large collection, it is incremental and grows continuously thanks to its users’ feedback. One important difference with regard to other datasets such as Torralba200880Recognition. () and Deng2009ImageNet:Database () is that, rather than employing images downloaded from the Internet, users take pictures specifically for object recognition. This signifies that MirBot images are focused on the target objects and gathered with minimum occlusions and plain backgrounds. Our team reviews the new images on a weekly basis in order to avoid inappropriate, unfocused or wrongly labeled samples, thus ensuring good quality data.
The MirBot dataset also includes a series of metadata that could be used to constrain the search space. These metadata, which are detailed in MirBot:System (), are extracted from the smartphone sensors (angle with regard to the horizontal, gyroscope, flash, GPS, etc.), and have reverse geocoding information (type of place, country, closest points of interest, etc.) and EXIF camera data (aperture, brightness, ISO, etc.). Although all the images are stored with their associated metadata, the similarity search is currently performed using only visual information, although we hope to include metadata in the near future.
2 User interaction
As stated in Lew2006Content-basedChallenges (), beyond the one-shot queries in the early similarity-based search systems, the next generation of systems attempts to integrate continuous feedback from the user in order to learn more about the query. MirBot is designed as an interactive game and the objective of its interface is to minimize the number of user interactions.
In order to avoid the dispersion that can be caused when using free object identifiers, users can only assign class names selected from the WordNet ontology. The main advantage of WordNet is its semantic structure, which prevents ambiguities in the labels. WordNet synsets (synonym sets) are unique identifiers for meaningful semantic concepts, and each of them is linked to a definition, although they can be related to different lemmas. Similarly to ImageNet, we use the WordNet synsets as class identifiers.
In MirBot, only portrayable objects can be chosen from the WordNet hierarchy, including the following root categories: animals (lexicographer identifier: ), food/drinks (), plants (), and objects, which include both WordNet natural objects () and artifacts (). WordNet considers that an artifact is man-made whereas a natural object is not, but they have been merged in MirBot in order to simplify this concept for its users.
Figure 2 shows the block diagram of the user interface. Before sending the query image to the server, users can find settings in the app that allow to choose whether they wish to perform the classification by considering only their own images, so as to constrain the search space, or by using the whole dataset.
Once the query has been submitted, the classes of the most similar images are retrieved using the methodology described in Section 3, and the class of the first image in this ranking is given to the user for validation. If no images in the dataset are found to be similar, then an unknown object message is displayed and the user is asked to select a class from the WordNet dictionary.
In the case of the user confirming the server response, the image is labeled with that class and stored in the dataset. Otherwise, an alternative list containing the classes of the top- images in the ranking is shown. If none of them correspond to the actual class, the user can select a lemma (e.g., key) from the WordNet nouns, and then a definition related to this lemma (e.g., Metal device shaped...), which corresponds to the synset (class identifier).
The messages and the robot images in Show proposed class (see Figure 2) are dependent on the distance between the query image and the most likely prototype , in order to enhance the userâs experience.
Two constant thresholds are used for this purpose: , which indicates a certain answer (a very small distance from the query image to its closest prototype), and , which indicates that the object is unknown (a very large distance to its closest prototype).
If , then the message begins with I’m pretty sure that this is…. If , then the range is discretized into levels that trigger messages such as I’m almost sure that…, I have many doubts, but …, etc. Finally, if , then the user receives an answer indicating the object is unknown, as there are no prototypes in the database that are sufficiently close.
The messages and robot images in Show result similarly depend on the validated answer. If the proposed class is not correct, the closest common level in the WordNet hierarchy determines the system response (I was close, I was totally wrong, etc.).
Besides setting the WordNet class, users can also manage their images in My pics and label them using any text (e.g. My dog Toby). If the closest image in the ranking contains a label, it is also shown to the user in Show proposed class. Users can also learn more about their objects with Wikipedia and consult the WordNet hierarchy of a synset.
In the Android version, some gaming techniques were also implemented so as to involve more users. These additions include an IQ measure for the robots, a ranking of the best players and achievement badges.
After a photograph has been taken and sent to the server, object classification is performed in order to return an answer to the user. This section describes the image features and the similarity search techniques used to retrieve the most likely class. At present, only the image contents within the ROI are used to extract the visual descriptors in order to perform the classification.
3.1 Visual descriptors
Visual descriptors are extracted from the image on the server side. It is now also feasible to integrate feature extraction into a smartphone app mobilenets (). However, performing this on the server increases the flexibility of the system, as the descriptors and classifiers can easily be updated without having to modify the source code on the client side. Moreover, the incremental nature of the proposed method (new classes are continuously added) requires that the information must be centralized.
Traditional local invariant descriptors (Mikolajczyk2004ScaleDetectors, ), such as SIFT Lowe2004DistinctiveKeypoints () and SURF Bay2008Speeded-UpSURF (), and global descriptors such as color histograms (vandeSande2010EvaluatingRecognition, ; Jeong2001Histogram-BasedRetrieval, ) have been extensively used for object recognition in literature. Some methods also combined different features, such as color with local descriptors (Fernando2012DiscriminativeClassification, ), or shape with color and texture (Banerji2013NewClassification, ).
However, since 2013 most systems have relied on deep learning, as convolutional neural networks have dramatically improved the state of the art of visual object recognition. As pointed out in LeCun2015DeepLearning (), the ability of these techniques to learn the representations directly from the raw data have revolutionized many classification tasks related to image, video, speech and audio.
The visual features used in the initial MirBot version MirBot:System () were TopSURF local descriptors (Thomee2010TOP-SURF:Toolkit, ) and global YCbCr color histograms weighted with a two-dimensional Gaussian function. These features were replaced with neural codes in July 2015, as they clearly outperformed handcrafted descriptors. This section describes the different features evaluated using the MirBot dataset.
3.1.1 Local handcrafted descriptors
In MirBot:System () we initially used the TopSURF (Thomee2010TOP-SURF:Toolkit, ) toolkit to obtain a histogram of local descriptors for each image. This method calculates the Speed-Up Robust Features (SURF (Bay2008Speeded-UpSURF, ) interest points), and clusters them (Philbin2007ObjectMatching, ) into a bag of features so as to yield a dictionary of visual words. Those visual words that do not occur very frequently are emphasized using a tf-idf (Salton1983IntroductionRetrieval., ) weighting technique. Finally, the descriptor consists of a tf-idf histogram obtained by selecting those visual words with the highest score (the top words).
3.1.2 Global handcrafted descriptors
Standard SURF features do not consider color Jeong2001Histogram-BasedRetrieval (), which can be relevant when attempting to identify certain classes. The TopSURF descriptors in the initial version of MirBot MirBot:System () were, therefore, complemented with color histograms in order to improve accuracy.
Experiments with which to evaluate different color spaces (RGB, HSV, CIE-LUV and YCrCb) were carried out, and the best results were obtained using YCbCr, which has proven to be relatively robust to changes in lighting. The color value of each pixel was additionally weighted using a two-dimensional Gaussian function to obtain a weighted color histogram, as detailed in MirBot:System (). The goal of this weighting is to give less relevance to the colors that appear on the edges of the image and more weight to those on the middle, which is where the objects to be recognized are likely located, given their ROI.
3.1.3 Neural codes using pretrained models
Transfer learning is a common technique consisting of training a network using a given dataset and repurposing (or transferring) the features learned to a different dataset or network Yosinski2014How (). This process usually obtains remarkable results when the features are general (suitable for both base and target tasks).
We have evaluated several CNN models in order to obtain the neural codes:
AlexNet Krizhevsky2012ImageNetNetworks () is a CNN with 8 layers. Since it outperformed previous methods by a large margin in the ImageNet ILSVRC12 (Russakovsky2014ImageNetChallenge, ) challenge (which contains about 1.3 million images distributed in 1,000 object categories), most image recognition systems rely on CNNs.
GoogLeNet Szegedy2014GoingConvolutions () is a well-known CNN which achieved the best results in the ImageNet ILSVRC14 challenge. This network contains 22 layers but uses 12 times fewer parameters than AlexNet, and is based on a series of chained Inception modules.
VGG-16 and VGG-19 Simonyan14 (). VGG-16 has 13 convolutional and 3 fully-connected layers, whereas VGG-19 is composed of 16 convolutional and 3 fully-connected layers. Both topologies use Dropout and Max-pooling techniques and ReLU activation functions.
Inception21k is an Inception v2 network based on a GoogleNet with Batch Normalization Ioffe15:ArXiv (). We call this network Inception21k because it was trained with the full ImageNet dataset (14,197,087 images in 21,841 classes), unlike AlexNet and GoogleNet, which were trained with the 1,000 classes from the ILSVRC challenge.
Inception v3 inceptionv3 (). This architecture has 6 convolutional layers followed by 3 inception modules and a last fully connected layer. It has fewer parameters than other similar models thanks to the Inception modules whose design is based on two main ideas: The approximation of a sparse structure with spatially repeated dense components and the use of dimensionality reduction to maintain the computational complexity within bounds.
REsNet resnet (). The deep REsidual learning Network learns residual functions with reference to the layer inputs rather than learning unreferenced functions. This technique enables the use of a large number of layers. We have used the 50-layer version for our experiments.
Xception xception (). This model has 36 convolutional layers with a redesigned version of Inception modules which enable the depth-wise separable convolution operation. This architecture outperforms the Inception results using the same number of parameters.
We have used the AlexNet and GoogLeNet implementations from Caffe (Jia2014CaffeEmbedding, ) as a basis. These are almost identical to the original papers, with pretrained models from the ImageNet ILSVRC12 dataset. The original Inception21k network was trained by Mu Li for MXNet (MXNet, ), although we converted it to the Caffe framework by translating the MXNet Batch Normalization layers to Caffe BatchNorm layers with the learned mean and variance, followed by a Scale layer that applies the learned scale () and shift (). We have publicly released this model at https://github.com/pertusa/InceptionBN-21K-for-Caffe. In the case of the VGG, Xception, ResNet and Inception v3, we used the Keras keras () implementations with their default parameters.
The target images are forwarded through the pretrained network in order to obtain the visual descriptors, which are vectors containing the neuron output activations of the last hidden layer (excluding the output layer). For example, in the case of AlexNet, this is the Caffe layer fc7 with values, in GoogLeNet it is pool5/7x7_s1 with a dimension of , and in Inception21k it is the global_pool layer, which also has a dimension of .
The main advantages of using the neural codes with kNN rather than the output layer are that the model can be used incrementally, the output classes can be different to those used for training, and we also obtain a ranking of the most similar images.
3.2 Prototype matching
Given a sample query, the class of the most likely image among the set of prototypes is given to the user for validation. In this scenario, the techniques chosen for similarity search must be efficient and incremental owing to the real-time constraints of the proposed architecture, as users cannot wait long for a system response.
As efficiency is crucial for our system and the database is imbalanced (there are many more samples of some classes than of others He09 ()), an approximate nearest neighbor search technique has been chosen to match the image descriptors from a query image with the prototypes in the dataset.
This is done using the Spotify Annoy approximate kNN method annoy (). In the MirBot dataset, a good compromise between accuracy and performance was experimentally found by using an index of trees and setting the maximum number of nodes to . As MirBot should operate in real time, the index cannot be rebuilt (this takes about 1 minute) every time a new image is added. The descriptors that were not stored are, therefore, kept in the tree index in a database table, and the kNN is obtained by performing the search in both the Annoy indexed prototypes and in the descriptor table. The Annoy index is rebuilt weekly.
In our previous work MirBot:System (), the normalized cosine similarity was used to compare the TopSURF descriptors of two images, as in Thomee2010TOP-SURF:Toolkit (), whereas color histograms were compared using the Jensen-Shannon divergence (Lin1991DivergenceEntropy, ). Neural codes were compared using the Euclidean distance.
Given a query image , its nearest neighbor images from the set of prototypes are retrieved according to their distance . The classes of the top- images in this ranking with distances are then sent to the user for validation. As stated in the previous section, if no images in the dataset are found to be similar (), then an unknown object message is displayed and the user is asked to select a class from the WordNet dictionary.
In the initial MirBot system, TopSURF and color descriptors were also combined to yield a distance between two images:
where is the TopSURF distance, is the color distance, and is a parameter used to weight their contribution in the final distance .
The dataset is freely available on request through a web interface at http://www.mirbot.com/admin. Researchers can explore the images by means of an expandable tree view, and download them with their metadata and features .
Some examples of the dataset can be seen in Figure 3. Images that are out of focus, have a misplaced ROI, or are labeled with a wrong class can be marked as unreliable by the MirBot administrators, who can also delete samples with inappropriate contents.
As the data is dynamic, the statistics and success rate change over time. The following evaluation results refer to the MirBot dataset on October 23, 2016. On this date, users had added images distributed in classes. Figure 4 shows the temporal evolution of the data since MirBot has been available.
Since the dataset is user-driven, some objects appear more frequently than others and the classes are, therefore, imbalanced. Upon observing the WordNet root categories it will be noted that most images are objects (), followed by animals (), food/drinks (), and plants (). Figure 5 shows the top 40 classes by the number of samples.
The evaluation is performed using a 5-fold cross validation. Only the images belonging to the classes with more than one prototype were used in these experiments ( images from classes).
The accuracy is provided at two levels, top-1 and top-10. In top-1, a true positive is considered when the class of the closest prototype matches the query class. In top-10, a true positive is considered when the query class matches the class of any of its closest prototypes. Note that this does not mean that classes are considered, but rather the classes of the closest prototypes, and if they share the same class only one class is, therefore, taken into account.
4.2 Results using handcrafted descriptors
The MirBot base system introduced in MirBot:System () performed similarity search using only TopSURF descriptors and color histograms. The parameters were experimentally set to (color histogram size), (number of top words in TopSURF), and (TopSURF dictionary size). The evaluation results obtained using these values are shown in table 1.
Some experiments were performed by varying the size of the color histogram , the number of words , and the size of the dictionary . In all these configurations, the success rate varied by a maximum of , and these parameters do not, therefore, have a noticeable impact on the accuracy.
In MirBot:System (), a generic TopSURF dictionary of words was used. An experiment was also performed, which consisted of replacing this dictionary with one trained with the MirBot dataset in order to make the clusters more specific to the target data. The top-1 success rate increased from to when using the trained dictionary, as can be seen in table 1.
|Combined (w= 0.1)||36.61||55.95|
Surprisingly, the global color descriptor performed significantly better than the TopSURF local descriptor. The reason for this may be that some pictures from the same object have similar lighting conditions, as users tend to take them consecutively when the system returns a wrong answer.
We have also evaluated the combination of color and trained TopSURF descriptors (see Eq. 1), using the parameter to balance their contribution. Interestingly, although color outperforms TopSURF, the combination of both descriptors increases the individual success rate for some values of . The best results were specifically obtained with , which was significantly more successful than the color histograms, as can be seen in table 1.
4.3 Results using neural codes
Different parameters can be selected to extract the neural codes from an image. In particular, we have evaluated alternative CNN models, input image scaling techniques, neural activations from different inner layers, and the benefit of normalizing the neural codes using the norm before performing the kNN similarity search.
In our experiments we assessed that normalizing the neural codes using the norm consistently improved the accuracy, as can be seen in table 2. We also considered different image scaling methods (isotropic or anisotropic rescaling) and alternative input image sizes. As expected, the most adequate input size and rescaling parameters are the same as those originally used to train the networks. The neural codes were extracted from the last hidden layer, but we also carried out experiments using different layers. The last hidden layer obtained the best results in all cases and was, therefore, selected for the following experiments.
Neural codes from different topologies (detailed in Section 3.1.3) were also evaluated, and clearly outperformed the handcrafted features, as shown in table 1. MirBot initially used the combined Top-Surf and color descriptor. However, GoogLeNet replaced the handcrafted descriptors in June 2015, and from October 2016 to the present the features have been extracted with Inception21k, although given the evaluation results we plan to replace them with ResNet soon.
Table 1 shows the 5-fold cross validation average accuracy using the weights of the models trained with the ImageNet 1,000 classes subset (ILSVRC12), with the exception of Inception21k, which was trained with the full ImageNet dataset.
As the evaluation results show, the pretrained models obtain high success rates, especially when taking into account that the target classes are different from those used for training. The best accuracies are obtained with ResNet and Xception, which are those that yielded the highest accuracies in the ILSVRC contest.
Experiments were also performed by varying the value of . Table 2 shows the results obtained with and without the normalization of the neural codes. As can be seen, the highest accuracies are obtained with and normalized codes. This low value of could be owing to the highly imbalanced dataset, as some classes contain very few samples.
Besides transfer learning, we have also used the SoftMax output layer of the Inception21k network (namely Inception21k-direct), as this CNN was originally trained using the full ImageNet dataset and it therefore contains most of the classes that are present in MirBot. In these experiments we considered only the activations of the output neurons belonging to the MirBot classes that have more than one prototype. In this case, only images were used for evaluation, as of the MirBot classes were not present in the ImageNet classes. These results cannot be directly compared to the transfer learning approaches because the dataset is different, but we can obtain an approximate success rate.
The top-1 accuracy () in Inception21k-direct is much lower than that from the transfer learning representations, whereas the top-10 () is also lower but not to such a great extent. This discrepancy can be explained by the fact that some images may fit in different classes, signifying that these classes share a very similar content. For example, an image of a person may be classified either as a human-being, male, adult, or even homo-sapiens-sapiens.
4.4 Results with fine tuning and full training
Besides using the ImageNet pretrained weights, we have also trained the CNN models using the MirBot dataset. This training was performed using a maximum of 500 epochs, a mini-batch size of 16 samples, and an early stopping of 10 epochs. In the fine tuning experiments, the original fully-connected (FC) layers were replaced with three new FC layers in which the last one had the same number of neurons as the amount of MirBot classes (1,256) that contained more than one sample. We have evaluated three learning approaches:
Full training: The model is trained from scratch.
Fine tuning last layers: Fine-tuning using the weights of convolutional layers from ImageNet and replacing the last fully-connected layers with 3 fully-connected layers whose weights must be learned.
Fine tuning from the middle: Fine-tuning starting from the middle layer of the model whose weights were originally trained using ImageNet. The use of this scheme makes it possible to adjust more network weights for the new classes than in the previous approach.
The evaluation results obtained using the SoftMax activations of the trained models can be seen in Table 3. Full-training accuracy is very low because the number of samples in the MirBot dataset is not sufficient to train models that contain a large number of parameters and because the classes are highly imbalanced. The best results are obtained by fine-tuning only the last layers, although the accuracy is still below that obtained using the pretrained features and kNN.
In order to assess the accuracy in the same scenario employed with the pretrained models, experiments were also performed by replacing the SoftMax layer with a kNN after full training or fine tuning. This was done by extracting neural codes from the last hidden layer, normalizing them using , and then returning the class of the closest prototype () in the training set. The results are shown in Table 4. After testing different values, the size of the neural codes was set to 1,256, with the exception of the first column in this table (without training), in which they have the size of the last hidden layer from the original model.
It will be noted that, when fine-tuning the pretrained ILSVRC models starting from the middle layers and then use the normalized neural codes from the last hidden layer in the prediction stage, the accuracy outperforms the original pretrained models. Moreover, in all cases the accuracy obtained using the SoftMax output is clearly below that attained when using neural codes with kNN in the prediction stage.
4.5 Accuracy over time
We have also evaluated the handcrafted descriptors and the neural codes with regard to the number of images (see Figure 6). In this case, we performed top-1 leaving-one-out experiments with the first images that were added to the dataset, and then with the first images, until all the images had been considered. The goal of this experiment was to check the evolution of the MirBot success rate over time. For ease of viewing, only three representative CNN models pretrained with ILSVRC12 data are shown in Fig. 6.
The results of this experiment are very interesting. After the initial spikes owing to the small number of samples, the neural codes maintain an almost constant success rate despite the increasing number of classes. The main reason for this is that there are more classes over time but also more images in the database with which to compare them, thus signifying that the success rate over time remains constant with these descriptors. The accuracy for handcrafted descriptors, however, significantly decreased over time, showing that they are less robust.
4.6 Accuracy attained when increasing the minimum number of samples per class.
A similar experiment was performed with regard to the number of samples per class, during which the accuracy was evaluated using those classes which contain a minimum number of samples (from 1 to 100). The first column in figure 7 shows the top-1 accuracy using all classes, while the last one shows the accuracy attained when using only those classes with a minimum number of 100 samples. The experiments were performed with a 5-fold cross validation using the best parameters found in the previous evaluation (neural codes from the last hidden layer, norm and ).
It will be noted that the top-1 accuracy increases when more samples per class are available. As the number of new classes converges (as shown in Figure 6) and new images are continuously added, the accuracy is expected to increase over time as the MirBot dataset grows.
5 Conclusions and future work
This paper presents a description and evaluation of an interactive image recognition system for smartphones. The proposed method, which is freely available222http://www.mirbot.com for iOS and Android devices, allows a user to train an object recognition system using the classes from a semantic dictionary.
One of the main research results of this work is a multimodal dataset of images taken with smartphones and organized according to the WordNet ontology. This dataset is continuously growing as it is user-driven. Its images are labeled and stored with a series of metadata, within a ROI, with minimum occlusions, and usually with plain backgrounds, as they are specifically taken for this task and are not downloaded from the Internet. A web interface allows researchers to explore, review and download images, metadata and descriptors.
An interaction methodology is also presented, which allows users to easily set the image class in a few steps, adding dynamic messages in order to make the labeling process enjoyable. Besides setting the class of the images, users can also train the system to identify objects in their collections (such as stamps, butterflies, fossils, etc.) using the optional labels. However, most users treat the app as a casual game rather than a tool. Interestingly, our most frequent users are children, perhaps because they like the idea of teaching a robot.
Evaluation has been performed using top-1 and top-10 accuracy with different visual descriptors in order to obtain a baseline for future studies. The performance of color histograms, local features and different convolutional neural network representations is assessed.
The results obtained using handcrafted descriptors show that the combination of color and local descriptors may increase the success rate when compared to using them independently. However, CNN features outperform these descriptors by far, and their accuracy does not decrease with regard to the number of images when the database grows, unlike that which occurs with the handcrafted features. The usage of neural codes and kNN rather than the SoftMax output layer allows the system to be incremental when new classes are added.
As can be seen in the evaluation section, using ImageNet pretrained models to extract the neural codes yields a high accuracy even when the MirBot classes are different from those used to train the model. We also evaluated different fine-tuning strategies in order to adapt the weights to the data collected in MirBot. The best results were obtained using fine-tuning from the middle layers of a ResNet network, which was more accurate than the pretrained models. Moreover, we can conclude that the neural codes and kNN clearly provide more accuracy than the common prediction strategy (using the SoftMax output), as is shown in tables 3 and 4. We also verified that adding the norm improves the accuracy in all the scenarios evaluated.
As the dataset is incremental, the classes are imbalanced, which may be an issue for some classification methods (Barandela2003StrategiesProblems, ), although this can be overcome by employing sampling methods, such as using only those classes with a minimum number of prototypes. The success rate of MirBot does not decrease over time although the number of classes increases, and this occurs because the number of images per class also increases. The system is, therefore, continuously improving, as more classes are detected without a decrease in the success rate.
In summary, the main contributions of this work are: 1) An interactive workflow with which to perform object recognition and class validation; 2) A public dataset that is continuously growing with its usersâ feedback, and which contains labeled images and their associated metadata; 3) An analysis of different visual descriptors (handcrafted features and neural codes from different CNN topologies) in this dataset; 4) Different transfer learning methods evaluated for this task; 5) The conclusion that using normalized neural codes from the last hidden layer with kNN in the prediction stage provides a higher accuracy than the SoftMax output in all the cases evaluated.
Our future work includes the study of similarity metric learning techniques (Bellet2012SimilarityClassification, ) so as to replace Euclidean distance on kNN. Another work line is to include metadata for classification. These metadata can complement the visual descriptors with different multimodal techniques, such as early fusion, late fusion, or with a joint approach (using multimodal neural networks in a manner similar to that of Mao2014ExplainNetworks ()). It would also be possible to use a two-stage classification scheme (such as using metadata in a preprocessing stage to filter out certain image prototypes, or vice-versa).
The proposed dataset could also be analyzed using hierarchical classification methods. Although images from a root category are probably too different to benefit from hierarchical techniques, metadata could be effectively used as they may share common values in the case of certain subclasses.
This work was supported by the TIMUL project (TIN2013-48152-C2-1-R) and the University Institute for Computing Research (IUII) from the University of Alicante.
- (1) S. Bock, S. Newsome, Q. Wang, W. Zeng, X. Lin, J. Lu, iImage: An image based information retrieval application for the iPhone, in: 7th IEEE Consumer Communications and Networking Conference, CCNC, 2010, pp. 3–4. doi:10.1109/CCNC.2010.5421733.
A. G. Howard, M. Zhu, B. Chen, D. Kalenichenko, W. Wang, T. Weyand,
M. Andreetto, H. Adam, MobileNets:
Efficient Convolutional Neural Networks for Mobile Vision Applications,
- (3) K. Matusiak, P. Skulimowski, P. Strumillo, Object recognition in a mobile phone application for visually impaired users, in: IEEE 6th International Conference on Human System Interactions (HSI), 2013, pp. 479–484. doi:10.1109/HSI.2013.6577868.
- (4) C. Fellbaum, WordNet: An Electronic Lexical Database (1998). doi:10.1139/h11-025.
- (5) L. Barrington, D. Turnbull, G. Lanckriet, Game-powered Machine Learning., Proc. of the National Academy of Sciences of the United States of America 109 (17) (2012) 6411–6. doi:10.1073/pnas.1014748109.
- (6) L. von Ahn, L. Dabbish, Labeling images with a computer game, in: ACM Conference on Human Factors in Computing Systems, 2004, pp. 319 – 326. doi:10.1145/985692.985733.
- (7) A. Pertusa, A. J. Gallego, M. Bernabeu, Mirbot: A multimodal interactive image retrieval system, in: Pattern Recognition and Image Analysis, Vol. 7887 of Lecture Notes in Computer Science. 6th Iberian Conference, IbPRIA, 2013, pp. 197–204. doi:10.1007/978-3-642-38628-2.
- (8) A. Torralba, R. Fergus, W. T. Freeman, 80 Million Tiny Images: a Large Data Set for Nonparametric Object and Scene Recognition., IEEE Trans. on Pattern Analysis and Machine Intelligence (PAMI) 30 (11) (2008) 1958–1970. doi:10.1109/TPAMI.2008.128.
- (9) A. Khosla, T. Zhou, T. Malisiewicz, A. A. Efros, A. Torralba, Undoing the damage of dataset bias, Lecture Notes in Computer Science 7572 LNCS (part 1) (2012) 158–171. doi:10.1007/978-3-642-33718-5_12.
- (10) J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, L. Fei-Fei, ImageNet: A large-scale hierarchical image database, in: IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2009, pp. 2–9. doi:10.1109/CVPR.2009.5206848.
- (11) Y. LeCun, Y. Bengio, G. Hinton, Deep learning, Nature 521 (7553) (2015) 436–444. doi:10.1038/nature14539.
- (12) M. Lew, N. Sebe, C. Djeraba, R. Jain, Content-based multimedia information retrieval: State of the art and challenges, ACM Trans. on Multimedia Computing, Communications, and Applications 2 (1) (2006) 1–19. doi:10.1145/1126004.1126005.
- (13) K. Mikolajczyk, C. Schmid, Scale & affine invariant interest point detectors, International Journal of Computer Vision 60 (1) (2004) 63–86. doi:10.1023/B:VISI.0000027790.02288.f2.
- (14) D. G. Lowe, Distinctive image features from scale-invariant keypoints, International Journal of Computer Vision 60 (2) (2004) 91–110. doi:10.1023/B:VISI.0000029664.99615.94.
- (15) H. Bay, A. Ess, T. Tuytelaars, L. Van Gool, Speeded-Up Robust Features (SURF), Computer Vision and Image Understanding 110 (3) (2008) 346–359. doi:10.1016/j.cviu.2007.09.014.
- (16) K. E. A. van de Sande, T. Gevers, C. G. M. Snoek, Evaluating Color Descriptors for Object and Scene Recognition, IEEE Trans. on Pattern Analysis and Machine Intelligence (PAMI) 32 (9) (2010) 1582–1596. doi:10.1109/TPAMI.2009.154.
Color Image Retrieval, Tech. rep., Psych221/EE362 Project Report, Stanford
- (18) B. Fernando, E. Fromont, D. Muselet, M. Sebban, Discriminative feature fusion for image classification, in: IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 2012, pp. 3434–3441. doi:10.1109/cvpr.2012.6248084.
- (19) S. Banerji, A. Sinha, C. Liu, New image descriptors based on color, texture, shape, and wavelets for object and scene image classification, Neurocomputing 117 (2013) 173–185. doi:10.1016/j.neucom.2013.02.014.
- (20) B. Thomee, E. M. Bakker, M. S. Lew, TOP-SURF: A Visual Words Toolkit, ACM International Conference on Multimediadoi:10.1145/1873951.1874250.
- (21) J. Philbin, O. Chum, M. Isard, J. Sivic, A. Zisserman, Object retrieval with large vocabularies and fast spatial matching, in: IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 2007. doi:10.1109/CVPR.2007.383172.
- (22) G. Salton, M. J. McGill, Introduction to modern information retrieval, McGraw-Hill, Inc, New York, NY, USA, 1986. doi:10.1108/01435121111132365.
- (23) J. Yosinski, J. Clune, Y. Bengio, H. Lipson, How transferable are features in deep neural networks?, in: Proc. of Neural Information Processing Systems (NIPS), 2014. arXiv:1411.1792.
- (24) A. Krizhevsky, I. Sutskever, G. E. Hinton, ImageNet Classification with Deep Convolutional Neural Networks, in: Proc. Neural Information and Processing Systems (NIPS), 2012. arXiv:1102.0183.
- (25) O. Russakovsky, J. Deng, H. Su, J. Krause, S. Satheesh, S. Ma, Z. Huang, A. Karpathy, A. Khosla, M. Bernstein, A. C. Berg, L. Fei-Fei, ImageNet Large Scale Visual Recognition Challenge, International Journal of Computer Vision (IJCV)doi:10.1007/s11263-015-0816-y.
- (26) C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke, A. Rabinovich, Going Deeper with Convolutions, in: IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 2015. arXiv:1409.4842.
- (27) K. Simonyan, A. Zisserman, Very deep convolutional networks for large-scale image recognition, 2014. arXiv:1409.1556.
- (28) S. Ioffe, C. Szegedy, Batch normalization: Accelerating deep network training by reducing internal covariate shift, 2015. arXiv:1502.03167.
C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, Z. Wojna,
Rethinking the inception architecture
for computer vision, CoRR abs/1512.00567.
K. He, X. Zhang, S. Ren, J. Sun, Deep
residual learning for image recognition, CoRR abs/1512.03385.
F. Chollet, Xception: Deep learning with
depthwise separable convolutions, CoRR abs/1610.02357.
- (32) Y. Jia, E. Shelhamer, J. Donahue, S. Karayev, J. Long, R. Girshick, S. Guadarrama, T. Darrell, Caffe : Convolutional Architecture for Fast Feature Embedding, in: ACM Conference on Multimedia, 2014, pp. 675–678. doi:10.1145/2647868.2654889.
- (33) T. Chen, M. Li, N. Wang, M. a. Wang, Mxnet: A flexible and efficient machine learning library for heterogeneous distributed systems, 2016. arXiv:1512.01274.
- (34) F. Chollet, Keras, https://github.com/fchollet/keras (2015).
- (35) H. He, E. Garcia, Learning from imbalanced data, IEEE Trans. Knowl. Data Eng. 21 (2009) 1263–1284. doi:10.1109/TKDE.2008.239.
E. Bernhardsson, Annoy: Approximate
nearest neighbors in c++/python optimized for memory usage and loading/saving
to disk (2016).
- (37) J. Lin, Divergence measures based on the Shannon entropy, IEEE Trans. on Information Theory 37 (1) (1991) 145–151. doi:10.1109/18.61115.
- (38) R. Barandela, J.-S. Sánchez, V. García, E. Rangel, Strategies for learning in class imbalance problems, Pattern Recognition 36 (3) (2003) 849–851. doi:10.1016/S0031-3203(02)00257-1.
- (39) A. Bellet, A. Habrard, M. Sebban, Similarity Learning for Provably Accurate Sparse Linear Classification, in: 29th International Conference on Machine Learning, 2012. arXiv:1206.6476.
- (40) J. Mao, W. Xu, Y. Yang, J. Wang, A. L. Yuille, Explain Images with Multimodal Recurrent Neural Networks, in: NIPS Deep Learning Workshop, 2014, pp. 1–9. arXiv:1410.1090v1.