Large Scale Visual Recognition Challenge (ILSVRC) 2013: Classification spotlights
Additions to the ConvNet Image Classification Pipeline Andrew Howard – Andrew Howard Consulting Changes to Training: Use more pixels: Train on square patches from rectangular image instead of cropped central square Additional color manipulation of contrast, brightness, color balance used on training patches Changes to Testing: Make Predictions at different scales and different views which use all pixels Previous: Used 10 predictions (2 flips * 5 translations) This Submission: Used 90 predictions (2 flips * 5 translations * 3 scales * 3 views) The number of predictions can be reduced with no loss of accuracy with stagewise regression Higher Resolution Models: Use a fully trained model and fine tune on image patches from a higher resolution image This can be trained in about 1/3 the number of epochs Predictions on higher resolution images give complimentary predictions to the base model Final Vision System achieves 13.6% error and is made of 5 base models and 5 higher resolution models Structure is the same as last year with fully connected layers twice as large, which doesn’t add much value Use Patches From: Instead of Patches From: View 1:View 2:View 3:
Cognitive Psychology Inspired Image Classification using Deep Neural Network Kuiyuan Yang, Microsoft Research Yalong Bai, Harbin Institute of Technology Yong Rui, Microsoft Research CognitiveVision team
Our Classification Scheme Dog Cat French bulldog English setter Maltese dog Basic Category Classification Easy to distinguish Dog Classification Given a image, predict its basic category firstly. … Egyptian cat Siamese cat tiger cat Cat Classification dalmatian … Predict sub category CognitiveVision team
Caffe: Open-Sourcing Deep Learning Yangqing Jia, Trevor Darrell, UC Berkeley Convolutional Architecture for Fast Feature Extraction – Seamless switching between CPU and GPU – Fast computation (2.5ms / image with GPU) – Full training and testing capability – Reference ImageNet model available A framework to support multiple applications: Publicly available at Classification Embedding Detection Your next Application!
Experiments for large scale visual recognition Deep CNN (following Krizhevsky et al’12) We tried: + Low level features &spatial granularities Where did we fail? Television (0.18)Hair spray (0.18)Coffee mug (0.10) Flute (0.10) - TV vs. Screen, - Coffee mug vs. Cup, - Flute vs. Microphone, - … top 1 acc = Appliance and instrument are confusing for us, including