Fruits are the most vital items of global diets because of their rich nutritional value, thereby providing very high demand and agricultural revenues to the economy. Among the fruit crops, pomegranate is a valuable one due to its highest antioxidant potential. However, most crops of pomegranate suffer from diseases, which greatly reduce agricultural yield and productivity. Thus, along with the increasing demand of the fruit, early detection as well as classification of diseases will prove very crucial in boosting the yield and taking appropriate measures for prevention. We propose a segmentation-based model using deep learning in this paper to conduct disease identification in pomegranates The process begins with pre-processing images that is primarily an activity of cropping and resizing of the images, followed by enhanced Wiener filtering, which eliminates noise and enhances the clarity of the images The preprocessed images are then further segmented using a CA_YV5GC algorithm, (Channel Attentive YOLOv5-based Grab Cut), which isolates diseased regions from the images. Then the optimized ResNet-152 network is applied to acquire the fundamental features embedding the texture along with the shape characteristics which could identify ailments related symptoms. Coati Optimization is applied to choose the most dominant features in the lower dimensional representation of the extracted information for the classification of the disease. Ultimately, classification is performed using a Deep Capsule Canonical Auto-encoder (DC_CAENet) to classify the disease type with higher accuracy. Adaptive Osprey Optimization is used to optimize the parameters of the model. The existing methods are compared with that results proved this technique to be more accurate and efficient as compared to traditional techniques.

Aphasia therapy for Mandarin-speaking patients presents distinct challenges due to the language’s tonal characteristics and the presence of unforeseen vocal resonance, which reduces intelligibility and distorts tone contours. Current automatic speech feedback systems face challenges managing such distortions, especially in real-time and customized clinical contexts. This paper develops a novel framework, named graph-based adaptive acoustic feedback control (GA-AFC), that integrates graph neural networks (GNNs) with reinforcement learning (RL) to model and suppress articulation-resonance mismatches in aphasic speech in a dynamic manner. Unlike black-box automatic speech recognition (ASR) and traditional autoregressive models, GA-AFC constructs an articulation-resonance graph based on acoustic features such as harmonicity, pitch, energy, and Mel-frequency cepstral coefficients (MFCCs). The system utilizes GNN encoders to capture phoneme-tonal transitions and employs an RL policy to adapt acoustic feedback in real-time. Experimental evaluations on three benchmark Mandarin datasets, i.e., Common Voice (Mandarin), AISHELL-1, and HKUST, demonstrate that GA-AFC achieves substantial improvements in both fluency enhancement and recognition accuracy. In the context of aphasic speech, the model achieves an average word error reduction (WER) of 17.2% relative to Wav2Vec 2.0 and 30.1% relative to DeepSpeech, alongside a 14.8% improvement in tone classification accuracy on the HKUST corpus. Regarding resonance suppression, GA-AFC logs a spectral deviation of baseline systems by 28.6%, achieving a MOS score of 4.4 (±0.3) in subjective listening tests, which surpasses all comparative models. Moreover, the system demonstrates rapid convergence, with adaptation times of less than 20 s and feedback latencies of under 140 ms , making it suitable for real-time clinical use. The findings indicate that GA-AFC provides a responsive, adaptable, and clinically applicable framework for customizable speech feedback in Mandarin aphasia therapy, proposing a novel approach to tone- and resonance-sensitive neural interventions in speech rehabilitation.