Leaf disease of plants causes great loss in productivity of crops. So proper take care of plants is mandatory. Plants can be affected by various diseases. So Early diagnosis of leaf disease is a good practice. Computer vision-based classification of leaf disease can be a great way in diagnosing diseases early. Early detection of diseases can lead to better treatment. Vision based technology can identify disease quickly. Though deep learning is trending and using vastly for recognition task, but it needs very large dataset and also consumes much time. This paper introduced a method to classify leaf diseases using Gist and LBP (Local Binary Pattern) feature. These manual feature extraction process need less time. Combination of gist and LBP features shows significant result in classification of leaf diseases. Gist is used as global feature and LBP as local feature. Gist can describe an image very well as a scene. LBP is robust to illumination changes and occlusions and computationally simple. Various diseases of different plants are considered in this study. Gist and LBP features from images are extracted separately. Images are pre-processed before feature extraction. Then both feature matrix is combined using concatenation method. Training and testing is done on different plants separately. Different machine learning model is applied on the feature vector. Result from different machine learning algorithms is also compared. SVM performs better in classifying plant’s leaf dataset.

Classification of fish image is a complex issue in the field of pattern recognition. Fish classification is a complicated task. Physical shape, size, orientation etc. made it complex to classify. Selection of appropriate feature is also a great issue in image classification. Classification of fish image is very important in fishing service and agricultural field, fish industry, survey applications of fisheries and in other related area. For the assessment and counting of fishes, classification of fish image is also necessary as it can save time. This paper presents a fish image classification method with the robust Gist feature and Gray Level Co-occurrence Matrix (GLCM) feature. Noise removal and resizing of image is applied as pre-processing task. Gist and GLCM feature are combined to make a better feature matrix. Features are also tested separately. But combined feature vector performs better than individual. Classification is made on ten types of raw images of fish from two datasets -QUT and F4K dataset. The feature set is trained with different machine learning models. Among them, XgBoost performs with 90.2% and 98.08% accuracy for QUT and F4K dataset respectively.

Breast cancer is one of the most common cancer in women worldwide. Early detection of breast cancer can lead to better treatment and decrease in mortality. Mammogram image in medical technology, made it easier to analyze breast cancer. Mammography exam is a specialized imaging technique in medical to scan breast which results in mammogram image. Detecting breast cancer earlier, a patient can have several treatment options and also can save live. Early detection of breast cancer can leads to survive 93 percent or greater in the initial five years. This paper proposes a brseast cancer detection method from mammogram image sample by applying morphological operation on gray image rather than binary. Firstly, image is sent for gamma correction. Then it is converted to gray and applied morphological dilation. Again morphological opening operation is formed on the dilated image. Output of dilated and opening operation is then binarized. An AND operation is performed between both binary images. Some post processing like- small area filtering and hole filling task is took place. Then common unwanted object is removed. Finally rest of the region is the desired cancer infected region. Achieved performance is acceptable and satisfactory through the proposed method.

Skin lesion detection at early stage is very effective for patients. As a result, patients can get time for treatment. Moreover, this early detection helps a patient in the long-time survival. However, skin lesion detection from a dermoscopic images is not a general task. Due to inter and intra-observer variations in human interpretations, research on skin lesion detection from dermoscopic images become important. In this paper, we proposed a method to segment and detect lesion of skin from images. The proposed method is based on a set of rules of fuzzy logic approach. Firstly, a set of rules is applied on dermoscopic images. The output is then thresholded. Closing operation as a morphological tool is used on the thresholded image. Then area filtering takes a place which results in the desired output. With respect to different learning models, CNN shows better performance in classifying ISIC and Dermis-dermquest dataset. The system delivers a significant performance, which is remarkable and comparable.

Predicting human emotion from speech is now important research topic. One’s mental state can be understood by emotion. The proposed research work is emotion recognition from human speech. Proposed system plays significant role in recognizing emotion while someone is talking. It has a great use for smart home environment. One can understand the emotion of other who is in home or may be in other place. University, service center or hospital can get a valuable decision support system with this emotion prediction system. Features like-MFCC (Mel-Frequency Cepstral Coefficients) and LPC are extracted from audio sample signal. Audios are collected by recording speeches. A test also applied by combining self-collected dataset and popular Ravdees dataset. Self-collected dataset is named as ABEG. MFCC and LPC features are used in this study to train and test for predicting emotion. This study is made on angry, happy and neutral emotion classes. Different machine learning algorithms are applied here and result is compared with each other. Logistic regression performs well as compared to other ML algorithm.

Bengali text based news portal is now very common and increasing day by day. With easy access of internet technology, reading news through online is now a regular task. Different types of news are represented in the news portal. The system presented in this paper categorizes the news headline of news portal or sites. Prediction is made by machine learning algorithm. Large number of collected data are trained and tested. As pre-processing tasks such as tokenization, digit removal, removing punctuation marks, symbols, and deletion of stop words are processed. A set of stop words is also created manually. Strong stop words leads to better performance. Stop words deletion plays a lead role in feature selection. For optimization, genetic algorithm is used which results in reduced feature size. A comparison is also explored without optimization process. Dataset is established by collecting news headline from various Bengali news portal and sites. Resultant output shows well performance in categorization.

Flowers are blessing of nature. Classification of flowers as a natural image is difficult as they are surrounded by background. So a segmentation phase is needed to separate the flower from background as good as possible. Computer vision has gained much attention for classification task. This paper proposes a method to classify flower with the help of LBP and SURF as features and SVM as a classifier. Input image is pre-processed for enhancement of image quality. Then the image is segmented by applying active contour segmentation method. After segmentation of the image, LBP and SURF features are extracted. SURF features are extracted from MSER regions. Then both features are concatenated. These concatenated features are sent for classification to SVM classifier. Quadratic SVM is employed here. Quadratic SVM trains these feature and tests to classify. We also tried out with different classifier. But they provide poor results. Proposed quadratic SVM achieves an accuracy of 87.2% which is significant and comparable for this classification taskK.