Pona fish

The fish were photographed against a black background in daylight conditions to ensure consistency and clarity, making the dataset well-suited for machine learning applications. *Dataset Composition:* - *Good Samples (Healthy):* The dataset includes more than 250 images of healthy Rohu fish. These images depict fish with vibrant, unblemished scales, clear eyes, and overall robust physical condition. The good samples serve as positive examples for training the classification model, representing the desirable state of the fish. - *Bad Samples (Unhealthy):* The dataset also contains over 250 images of unhealthy Rohu fish.

These fish may show signs of disease, physical deformities, discoloration, or damage, indicating poor health. These images form the negative class, essential for teaching the model to recognize undesirable conditions. *Data Collection Setup:* Images were captured using the Xiaomi 11i mobile camera, chosen for its high-quality imaging capabilities. The use of a black background was a strategic choice to create a stark contrast with the fish, highlighting the fish's features and reducing distractions from the surrounding environment.

Daylight conditions were used to maintain consistent and natural lighting, which is crucial for capturing the true color and texture of the fish. *Image Characteristics:* The images in the dataset vary in terms of fish size, coloration, and health status, providing a comprehensive representation of Rohu fish under different conditions. This diversity ensures that the classification model can generalize well across different scenarios and accurately identify the health status of the fish. *Data Annotation:* Each image in the dataset is carefully annotated as either "good" or "bad" based on the health condition of the fish.

These annotations serve as the ground truth for training the machine learning model, ensuring that it learns to differentiate between healthy and unhealthy fish accurately. *Data Preprocessing:* Before feeding the data into the model, several preprocessing steps are applied: - *Resizing:* Images are resized to a standard dimension to ensure across the dataset. - *Normalization:* Pixel values are normalized to bring consistency and enhance the learning process of the model. - *Data Augmentation:* Techniques such as rotation, flipping, and scaling are applied to the images to increase the dataset’s variability, improving the model's ability to generalize to new, unseen data.