Learning Binary Codes and Binary Weights for Efficient Classification
نویسندگان
چکیده
This paper proposes a generic formulation that significantly expedites the training and deployment of image classification models, particularly under the scenarios of many image categories and high feature dimensions. As a defining property, our method represents both the images and learned classifiers using binary hash codes, which are simultaneously learned from the training data. Classifying an image thereby reduces to computing the Hamming distance between the binary codes of the image and classifiers and selecting the class with minimal Hamming distance. Conventionally, compact hash codes are primarily used for accelerating image search. Our work is first of its kind to represent classifiers using binary codes. Specifically, we formulate multi-class image classification as an optimization problem over binary variables. The optimization alternatively proceeds over the binary classifiers and image hash codes. Profiting from the special property of binary codes, we show that the sub-problems can be efficiently solved through either a binary quadratic program (BQP) or linear program. In particular, for attacking the BQP problem, we propose a novel bit-flipping procedure which enjoys high efficacy and local optimality guarantee. Our formulation supports a large family of empirical loss functions and is here instantiated by exponential / hinge losses. Comprehensive evaluations are conducted on several representative image benchmarks. The experiments consistently observe reduced complexities of model training and deployment, without sacrifice of accuracies.
منابع مشابه
Deep Error-correcting Output Codes
Existing deep networks are generally initialized with unsupervised methods, such as random assignments and greedy layerwise pre-training. This may result in the whole training process (initialization/pre-training + fine-tuning) to be very timeconsuming. In this paper, we combine the ideas of ensemble learning and deep learning, and present a novel deep learning framework called deep error-corre...
متن کاملFeature-based Malicious URL and Attack Type Detection Using Multi-class Classification
Nowadays, malicious URLs are the common threat to the businesses, social networks, net-banking etc. Existing approaches have focused on binary detection i.e. either the URL is malicious or benign. Very few literature is found which focused on the detection of malicious URLs and their attack types. Hence, it becomes necessary to know the attack type and adopt an effective countermeasure. This pa...
متن کاملLearning Binary Code Representations for Effective and Efficient Image Retrieval
Title of dissertation: LEARNING BINARY CODE REPRESENTATIONS FOR EFFECTIVE AND EFFICIENT IMAGE RETRIEVAL Bahadir Ozdemir, Doctor of Philosophy, 2016 Dissertation directed by: Professor Larry S. Davis Department of Computer Science The size of online image datasets is constantly increasing. Considering an image dataset with millions of images, image retrieval becomes a seemingly intractable probl...
متن کاملتغییرات جدید الگوی دودویی محلی و طبقه بندی و قسمت بندی تصاویر بافتی بستر دریا
Texture analysis plays an important role in image processing. Considering the extraordinary appearance texture sonar images, texture analysis are good choices for analysis of acoustic seabed images. Local binary pattern (LBP) operator is a very efficient and multi-resolution texture descriptor. It acquires appropriate information from the illumination and moods of images. Despite many developin...
متن کاملAutomatic Face Recognition via Local Directional Patterns
Automatic facial recognition has many potential applications in different areas of humancomputer interaction. However, they are not yet fully realized due to the lack of an effectivefacial feature descriptor. In this paper, we present a new appearance based feature descriptor,the local directional pattern (LDP), to represent facial geometry and analyze its performance inrecognition. An LDP feat...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- CoRR
دوره abs/1603.04116 شماره
صفحات -
تاریخ انتشار 2016