IJRCIT- International Journal of Research in Computer & Information Technology

Archives of International Journal of Research in Computer & Information Technology(IJRCIT)

Volume 6 Issue 4 September 2021

1. Study of Plant Diseases Detection & Classification Using Image Processing Techniques

AUTHOR NAME : Ketaki A. Kadu, Prof. K. B. Bijwe

ABSTRACT : Agriculture is one of the important sources of livelihood in the world. Nowadays the agriculture sector also providing employment opportunities to the village people on large scale in developing countries like India. In India, many crops are cultivating and according to the survey, nearly 75% population is dependent on agriculture. Most of the Indian farmers are adopting manual cultivation due to a lack of technical knowledge and are unaware of what kind of crops grows well on their land. When plants are affected by a number of diseases through their leaves that will directly affect the production of agriculture and profitable loss with the reduction in quality of crops. So that healthy plant leaves are very important for the fast growth of plants and to increase the production of crops. Identification of proper diseases in plants leaves is challenging for farmers and also for researchers. Nowadays farmers are spraying pesticides on the plants without knowing the proper disease of the plant and it also affects humans directly or indirectly by health and economically also. Therefore, to detect these plant diseases many techniques need to be adopted. In this paper, we have done a survey on different plants disease and various available advanced techniques to detect proper plant diseases.

Download

2. Hybrid PSO-CNN-LSTM Model for Accurate Multi-Class Skin Disease Classification from Dermoscopic Images

AUTHOR NAME : Ranjana Gawai

ABSTRACT : Skin disease detection, particularly for conditions such as melanoma and other pigmented lesions, is a critical task in medical diagnosis due to its direct impact on early treatment and patient survival. Traditional diagnostic methods rely heavily on expert interpretation, which can be subjective and time-consuming. To address these challenges, this study proposes a hybrid deep learning framework that integrates Convolutional Neural Networks, Long Short-Term Memory, and Particle Swarm Optimization for accurate and efficient skin disease classification. The CNN component is employed to extract discriminative spatial features from dermoscopic images, while the LSTM network captures sequential dependencies among the extracted features. Furthermore, PSO is utilized for optimal feature selection to reduce redundancy and enhance model performance. The proposed model is evaluated on the HAM10000 dataset, which consists of 10,015 dermoscopic images across seven skin disease classes. Experimental results demonstrate that the proposed PSO-CNN+LSTM model achieves superior performance compared to baseline models, attaining an accuracy of 98.2%, precision of 97.8%, recall of 97.3%, and F1-score of 97.5%. The findings highlight the effectiveness of combining deep learning with optimization techniques to improve classification accuracy and computational efficiency. The proposed framework provides a robust and reliable solution for automated skin disease detection and has significant potential for real-world clinical applications

Download