Abstract
One of the challenges in fine-grained classification is that subcategories with significant similarity are hard to be distinguished due to the equal treatment of all subcategories in existing algorithms. In order to solve this problem, a fine-grained image classification method by combining a bilinear convolutional neural network (B-CNN) and the measurement of subcategory similarities is proposed. Firstly, an improved weakly supervised localization method is designed to obtain the bounding box of the main object, which allows the model to eliminate the influence of background noise and obtain more accurate features. Then, sample features in the training set are computed by B-CNN so that the fuzzing similarity matrix for measuring interclass similarities can be obtained. To further improve classification accuracy, the loss function is designed by weighting triplet loss and softmax loss. Extensive experiments implemented on two benchmarks datasets, Stanford Cars-196 and Caltech-UCSD Birds-200-2011 (CUB-200-2011), show that the newly proposed method outperforms in accuracy several state-of-the-art weakly supervised classification models.
Highlights
Fine-grained image classification is a challenging task in computer vision
Inspired by the work of SCDA [21], we propose to add weakly supervised localization into a traditional bilinear convolutional neural network (B-convolutional neural networks (CNNs)), which can eliminate the influence of background noise and extract features more accurately
The problem is that the background of the original image brings noise to feature extraction, so a localization method based on a B-CNN model is proposed to locate the main object and remove the background under the weakly supervised setting
Summary
Fine-grained image classification is a challenging task in computer vision. Inspired by past developments in handcrafted features, many CNN-based fine-grained classification approaches have been proposed that benefit a wide variety of application scenarios in both industry and research, such as image retrieval, wildlife protection, and medical-image analysis [6]. The main challenge of fine-grained classification is that the differences between different subcategories are usually subtle and local, so how to locate discriminative regions has become a hot topic. For fine-grained image classification, similarities between different subcategories are different. Focused on the issue that existing algorithms treat all subcategories by equal cost, which limits the classification ability of subcategories with significant similarity, a bilinear CNN fine-grained image-classification method based on subcategory similarity is proposed.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
