Learning Geometric Equivalence between Patterns Using Embedding Neural Networks

Despite impressive results in object classification, verification and recognition, most deep neural network based recognition systems become brittle when the view point of the camera changes dramatically. Robustness to geometric transformations is highly desirable for applications like wild life monitoring where there is no control on the pose of the objects of interest. The images of different objects viewed from various observation points define equivalence classes where by definition two images are said to be equivalent if they are views from the same object. These equivalence classes can be learned via embeddings that map the input images to vectors of real numbers. During training, equivalent images are mapped to vectors that get pulled closer together, whereas if the images are not equivalent their associated vectors get pulled apart. In this work, we present an effective deep neural network model for learning the homographic equivalence between patterns. The long term aim of this research is to develop more robust manta ray recognizers. Manta rays bear unique natural spot patterns on their bellies. Visual identification based on these patterns from underwater images enables a better understanding of habitat use by monitoring individuals within populations. We test our model on a dataset of artificially generated patterns that resemble natural patterning. Our experiments demonstrate that the proposed architecture is able to discriminate between patterns subjected to large homographic transformations.