CNN Features Off-the-Shelf: An Astounding Baseline for Recognition (2014-03-23T00:00:00.000000Z)

TL;DR

A series of experiments conducted for different recognition tasks using the publicly available code and model of the OverFeat network which was trained to perform object classification on ILSVRC13 suggest that features obtained from deep learning with convolutional nets should be the primary candidate in most visual recognition tasks.

Abstract

Recent results indicate that the generic descriptors extracted from the convolutional neural networks are very powerful. This paper adds to the mounting evidence that this is indeed the case. We report on a series of experiments conducted for different recognition tasks using the publicly available code and model of the OverFeat network which was trained to perform object classification on ILSVRC13. We use features extracted from the OverFeat network as a generic image representation to tackle the diverse range of recognition tasks of object image classification, scene recognition, fine grained recognition, attribute detection and image retrieval applied to a diverse set of datasets. We selected these tasks and datasets as they gradually move further away from the original task and data the OverFeat network was trained to solve. Astonishingly, we report consistent superior results compared to the highly tuned state-of-the-art systems in all the visual classification tasks on various datasets. For instance retrieval it consistently outperforms low memory footprint methods except for sculptures dataset. The results are achieved using a linear SVM classifier (or L2 distance in case of retrieval) applied to a feature representation of size 4096 extracted from a layer in the net. The representations are further modified using simple augmentation techniques e.g. jittering. The results strongly suggest that features obtained from deep learning with convolutional nets should be the primary candidate in most visual recognition tasks.

Authors

A. Razavian

1 papers

Hossein Azizpour

2 papers

Josephine Sullivan

1 papers

TL;DR

Abstract

Authors

References58 items

ImageNet Large Scale Visual Recognition Challenge

Learning and Transferring Mid-level Image Representations Using Convolutional Neural Networks

DeepFace: Closing the Gap to Human-Level Performance in Face Verification

Multi-scale Orderless Pooling of Deep Convolutional Activation Features

Mining Mid-level Features for Image Classification

OverFeat: Integrated Recognition, Localization and Detection using Convolutional Networks

DeepPose: Human Pose Estimation via Deep Neural Networks

Mid-level Visual Element Discovery as Discriminative Mode Seeking

Deformable Part Descriptors for Fine-Grained Recognition and Attribute Prediction

Learning Discriminative Part Detectors for Image Classification and Cosegmentation

To Aggregate or Not to aggregate: Selective Match Kernels for Image Search

PANDA: Pose Aligned Networks for Deep Attribute Modeling

Visualizing and Understanding Convolutional Networks

Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation

DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition

Oriented pooling for dense and non-dense rotation-invariant features

POOF: Part-Based One-vs.-One Features for Fine-Grained Categorization, Face Verification, and Attribute Estimation

Efficient Object Detection and Segmentation for Fine-Grained Recognition

Harvesting Mid-level Visual Concepts from Large-Scale Internet Images

Subcategory-Aware Object Classification

Blocks That Shout: Distinctive Parts for Scene Classification

All About VLAD

ImageNet classification with deep convolutional neural networks

Aggregating Local Image Descriptors into Compact Codes

Pose pooling kernels for sub-category recognition

Reconfigurable models for scene recognition

Three things everyone should know to improve object retrieval

Hierarchical matching with side information for image classification

A database for fine grained activity detection of cooking activities

Asymmetric hamming embedding: taking the best of our bits for large scale image search

Smooth object retrieval using a bag of boundaries

Scene recognition and weakly supervised object localization with deformable part-based models

BiCoS: A Bi-level co-segmentation method for image classification

Describing people: A poselet-based approach to attribute classification

The Caltech-UCSD Birds-200-2011 Dataset

Combining randomization and discrimination for fine-grained image categorization

Contextualizing object detection and classification

LIBSVM: A library for support vector machines

Object Bank: A High-Level Image Representation for Scene Classification & Semantic Feature Sparsification

Sparse Representations and Distance Learning for Attribute Based Category Recognition

A Discriminative Latent Model of Object Classes and Attributes

Learning a Fine Vocabulary

Large-scale image retrieval with compressed Fisher vectors

Describing objects by their attributes

Recognizing indoor scenes

Automated Flower Classification over a Large Number of Classes

Et al

Hamming Embedding and Weak Geometric Consistency for Large Scale Image Search

Lost in quantization: Improving particular object retrieval in large scale image databases

Object retrieval with large vocabularies and fast spatial matching

Scalable Recognition with a Vocabulary Tree

Care of aged doctors

Negative evidences and co-occurrences in image retrieval: the benet of PCA and whitening

The Pascal Visual Object Classes Challenge 2006 ( VOC 2006 ) Results

and as an in

Attribute-Based Classification for Zero-Shot Visual Object Categorization

Svmstruct support vector machine for complex outputs

: Large

Field of Study

Journal Information

Name

Page