In this article, the overview of the following research works are described. (1) Multi-spectral image analysis: discrimination of changed and unchanged pixels in the multi-temporal satellite images using the unsupervised techniques like Otsu’s thresholding of local neighborhood information based technique; local binary similarity pattern descriptor based technique, (2) Hand gesture recognition: static hand gesture recognition in real-time; design of robust classifier against orientation and scale sensitivity; development of an interactive system for hand gesture based applications, (3) Electroencephalogram signal analysis for detection of P300 based brain computer interface spellers with alphabetic writing systems: using principal component analysis and ensemble of weighted support vector machines (SVMs); using score normalization of ensemble SVMs, (4) Electrocardiogram signal analysis: cardiac arrhythmia classification based on convolutional neural network.

中文翻译：

---

多光谱成像，手势识别，EEG和ECG信号处理的研究工作概述

在本文中，描述了以下研究工作的概述。（1）多光谱图像分析：使用大津基于局部邻域信息的阈值化技术等无监督技术，对多时相卫星图像中变化和不变的像素进行识别；基于局部二进制相似度模式描述符的技术，（2）手势识别：静态手势实时识别；针对方向和比例灵敏度的鲁棒分类器设计；开发基于手势的交互系统，（3）脑电信号分析，用于使用字母书写系统检测基于P300的脑计算机接口拼写器：使用主成分分析和加权支持向量机（SVM）集成；使用整体SVM的分数归一化，