Contextual image classification

About: Contextual image classification is a research topic. Over the lifetime, 33006 publications have been published within this topic receiving 931963 citations.

The Bag-of-Words Method with Different Types of Image Features and Dictionary Analysis

Abstract: Algorithms from the field of computer vision are widely applied in various fields including security, monitoring, automation elements, but also in multimodal human-computer interactions where they are used for face detection, body tracking and object recognition. Designing algorithms to reliably perform these tasks with limited computing resources and the ability to detect the presence of nearby people and objects in the background, changes in illumination and camera pose is a huge challenge for the field. Many of these problems use different classification methods. One of many image classification algorithms is Bag-of-Words (BoW). Originally, the classic BoW algorithm was used mainly for the natural language, so its direct application to computer vision issues may not be effective enough. The algorithm presented in this article contains a number of modifications that facilitate application of many types of characteristic features extracted from an image, image representation analysis and an adaptive clustering algorithm to create a dictionary of image features. These modifications affect classification result, which was confirmed in the experimental research.

Learning CNN Filters From User-Drawn Image Markers for Coconut-Tree Image Classification

Abstract: Identifying species of trees in aerial images is essential for land-use classification, plantation monitoring, and impact assessment of natural disasters. The manual identification of trees in aerial images is tedious, costly, and error-prone, so automatic classification methods are necessary. Convolutional neural network (CNN) models have well succeeded in image classification applications from different domains. However, CNN models usually require intensive manual annotation to create large training sets. One may conceptually divide a CNN into convolutional layers for feature extraction and fully connected layers for feature space reduction and classification. We present a method that needs a minimal set of user-selected images to train the CNN's feature extractor, reducing the number of required images to train the fully connected layers. The method learns the filters of each convolutional layer from user-drawn markers in image regions that discriminate classes, allowing better user control and understanding of the training process. It does not rely on optimization based on backpropagation, and we demonstrate its advantages on the binary classification of coconut-tree aerial images against one of the most popular CNN models.

An Active Learning Approach for Multi-Label Image Classification with Sample Noise

Abstract: Multi-label active learning for image classification has been a popular research topic. It faces several challenges, even though related work has made great progress. Existing studies on multi-label active learning do not pay attention to the cleanness of sample data. In reality, data are easily polluted by external influences that are likely to disturb the exploration of data space and have a negative effect on model training. Previous methods of label correlation mining, which are purely based on observed label distribution, are defective. Apart from neglecting noise influence, they also cannot acquire sufficient relevant information. In fact, they neglect inner relation mapping from example space to label space, which is an implicit way of modeling label relationships. To solve these issues, we develop a novel multi-label active learning with low-rank application (ENMAL) algorithm in this paper. A low-rank model is constructed to quantize noise level, and the example-label pairs that contain less noise are emphasized when sampling. A low-rank mapping matrix is learned to signify the mapping relation of a multi-label domain to capture a more comprehensive and reasonable label correlation. Integrating label correlation with uncertainty and considering sample noise, an efficient sampling strategy is developed. We extend ENMAL with automatic labeling (denoted as AL-ENMAL) to further reduce the annotation workload of active learning. Empirical research demonstrates the efficacy of our approaches.

Attention neural network for water image classification under IoT environment

Abstract: With significant development of sensors and Internet of things (IoT), researchers nowadays can easily know what happens in water ecosystem by acquiring water images. Essentially, growing data category and size greatly contribute to solving water pollution problems. In this paper, we focus on classifying water images to sub-categories of clean and polluted water, thus promoting instant feedback of a water pollution monitoring system that utilizes IoT technology to capture water image. Due to low inter-class and high intra-class differences of captured water images, water image classification is challenging. Inspired by the ability to extract highly distinguish features of Convolutional Neural Network (CNN), we aim to construct an attention neural network for IoT captured water images classification that appropriately encodes channel-wise and multi-layer properties to accomplish feature representation enhancement. During construction, we firstly propose channel-wise attention gate structure and then utilize it to construct a hierarchical attention neural network in local and global sense. We carried out comparative experiments on an image dataset about water surface with several studies, which showed the effectiveness of the proposed attention neural network for water image classification. We applied the proposed neural network as a key part of a water image based pollution monitoring system, which helps users to monitor water pollution breaks in real-time and take instant actions to deal with pollution.

Radon cumulative distribution transform subspace modeling for image classification

Abstract: We present a new supervised image classification method applicable to a broad class of image deformation models. The method makes use of the previously described Radon Cumulative Distribution Transform (R-CDT) for image data, whose mathematical properties are exploited to express the image data in a form that is more suitable for machine learning. While certain operations such as translation, scaling, and higher-order transformations are challenging to model in native image space, we show the R-CDT can capture some of these variations and thus render the associated image classification problems easier to solve. The method -- utilizing a nearest-subspace algorithm in R-CDT space -- is simple to implement, non-iterative, has no hyper-parameters to tune, is computationally efficient, label efficient, and provides competitive accuracies to state-of-the-art neural networks for many types of classification problems. In addition to the test accuracy performances, we show improvements (with respect to neural network-based methods) in terms of computational efficiency (it can be implemented without the use of GPUs), number of training samples needed for training, as well as out-of-distribution generalization. The Python code for reproducing our results is available at this https URL.