Data-driven and task-driven image classification method

Abstract

The invention discloses a data-driven and task-driven image classification method. The data-driven and task-driven classification method comprises the steps that a convolutional neural network structure is designed according to the scale of data sets and image content; a convolutional neural network model is trained through the given classified data sets; feature expression is extracted from training set images through a trained convolution neural network; images to be tested are input into the trained convolutional neural network and are classified. The data-driven and task-driven image classification method is based on nonlinear convolution feature learning, and the model can be adapted to the data sets through a date driving mode, so that the specific data set can be better described; errors of K-nearest neighbors can be directly optimized through a task-driving mode, and therefore a better performance can be obtained with respect to a K-nearest neighbor task; efficient training can be conducted through a GPU in the training stage, and efficient K-nearest neighbor image classification can be achieved just through a CPU in the testing stage; in this way, the data-driven and task-driven image classification method is quite suitable for a large-scale image classification task, a retrieval task and the like.

Description

technical field [0001] The invention relates to the technical field of image classification in computer vision, in particular to a data and task-driven image classification method. Background technique [0002] Image classification is one of the most basic research problems in computer vision. The problem to be solved is to automatically judge whether a certain type of object is contained in a given image. Image classification is a core topic of vision research. Many other vision researches rely on and involve image classification problems, such as object detection and tracking in images, image segmentation, object classification, detection and tracking in videos, behavior analysis, gesture recognition, etc. [0003] K-nearest neighbor image classification is an image classification method, which means that the K-nearest neighbor voting method is used when classifying images, that is, the category that appears most often in the K nearest images is predicted as the category o...

AI Technical Summary

Problems solved by technology

The bag-of-words model usually consists of the steps of low-level feature description, visual word generation, low-level feature encoding, feature aggregation, classifier training and testing. Before classifier training, we can think that the bag-of-words model uses an unsupervised For expression, whether it is the traditional low-level features such as SIFT and HOG or the middle-level feature expression of the bag-of-words model, the label information of the image is not used. Therefore, such feature expression is usually not the best for a parameterless model such as K-nearest neighbor classification. excellent

Images