Bird Species Categorization Using Pose Normalized Deep Convolutional Nets

Abstract

We propose an architecture for fine-grained visual categorization that approaches expert human performance in the classification of bird species. Our architecture first computes an estimate of the object's pose; this is used to compute local image features which are, in turn, used for classification. The features are computed by applying deep convolutional nets to image patches that are located and normalized by the pose. We perform an empirical study of a number of pose normalization schemes, including an investigation of higher order geometric warping functions. We propose a novel graph-based clustering algorithm for learning a compact pose normalization space. We perform a detailed investigation of state-of-the-art deep convolutional feature implementations and fine-tuning feature learning for fine-grained classification. We observe that a model that integrates lower-level feature layers with pose-normalized extraction routines and higher-level feature layers with unaligned image features works best. Our experiments advance state-of-the-art performance on bird species recognition, with a large improvement of correct classification rates over previous methods (75% vs. 55-65%).

Cited by in corpus (33)

• A comparison of dense region detectors for image search and fine-grained classification
• Looking for the Devil in the Details: Learning Trilinear Attention Sampling Network for Fine-grained Image Recognition
• Active and Continuous Exploration with Deep Neural Networks and Expected Model Output Changes
• Part Detector Discovery in Deep Convolutional Neural Networks
• Fine-grained Recognition Datasets for Biodiversity Analysis
• Re-rank Coarse Classification with Local Region Enhanced Features for Fine-Grained Image Recognition
• Enhance the Motion Cues for Face Anti-Spoofing using CNN-LSTM Architecture
• Crime Mapping from Satellite Imagery via Deep Learning
• The Application of Two-level Attention Models in Deep Convolutional Neural Network for Fine-grained Image Classification
• Actions and Attributes from Wholes and Parts
• Joint Multi-Person Pose Estimation and Semantic Part Segmentation
• DeepKSPD: Learning Kernel-matrix-based SPD Representation for Fine-grained Image Recognition
• Fine-grained Recognition in the Wild: A Multi-Task Domain Adaptation Approach
• Deep Quantization: Encoding Convolutional Activations with Deep Generative Model
• An Out-of-the-box Full-network Embedding for Convolutional Neural Networks
• Where to Focus: Deep Attention-based Spatially Recurrent Bilinear Networks for Fine-Grained Visual Recognition
• Building Graph Representations of Deep Vector Embeddings
• Real Time Fine-Grained Categorization with Accuracy and Interpretability
• Your "Flamingo" is My "Bird": Fine-Grained, or Not
• Data-driven Meta-set Based Fine-Grained Visual Classification
• RAMS-Trans: Recurrent Attention Multi-scale Transformer forFine-grained Image Recognition
• Learning a Discriminant Latent Space with Neural Discriminant Analysis
• One-Shot Fine-Grained Instance Retrieval
• DAF-NET: a saliency based weakly supervised method of dual attention fusion for fine-grained image classification
• Associating Multi-Scale Receptive Fields for Fine-grained Recognition
• Fine-Grained Visual Classification via Simultaneously Learning of Multi-regional Multi-grained Features
• Financial ticket intelligent recognition system based on deep learning
• Vision Recognition using Discriminant Sparse Optimization Learning
• Scalable Annotation of Fine-Grained Categories Without Experts
• Subset Feature Learning for Fine-Grained Category Classification
• Affine Non-negative Collaborative Representation Based Pattern Classification
• Weakly Supervised Keypoint Discovery
• Salvage Reusable Samples from Noisy Data for Robust Learning