Halmar Halide, Muhlis, Siti Fathana Azh Zahra
This study demonstrates how DeepSeek generates scripts for automated image processing, particularly for object detection and area comparison in noisy environments. The script processes two grayscale images containing irregularly shaped spots by adding noise to simulate real-world distortions. A median filter is applied to reduce noise, followed by Otsu’s thresholding for segmentation. The object boundaries are detected and outlined with red contours, making it easy to identify their shapes. The script then calculates the area of each object and compares them visually. It highlights the differences between the objects by applying a blue mask, showcasing the discrepancies. The results show that DeepSeek is effective in handling noise reduction, segmentation, contour detection, and shape analysis, proving its usefulness for automated image processing and object comparison tasks.
Keyword: DeepSeek, Automated image processing, Object detection, Noise reduction, Segmentation
