Comparison of image enhancement methods for pratima theft detection using artificial intelligence

The theft of pratima in Balinese temples threatens the spiritual and cultural balance of the community. These sacred objects, regarded as manifestations of God in Hinduism, hold profound religious significance, and their loss represents both material and spiritual desecration. To address this issue, this study investigates a security system that leverages image enhancement for low-light detection. Four techniques—contrast limited adaptive histogram equalization (CLAHE), adaptive histogram equalization (AHE), histogram equalization (HE), and gamma correction—were evaluated to improve image quality. CLAHE yielded the lowest mean squared error (MSE) of 21.16 and the highest peak signal-to-noise ratio (PSNR) of 38.13 dB. For object detection, VGG-19 and AlexNet were assessed. The best configuration, VGG-19 with HE, reached 83.33% accuracy and 93.75% recall, and achieved a receiver operating characteristic area under the curve (ROC AUC) of 0.90±0.02 across five runs. Thresholds derived from the ROC analysis were selected using the Youden J statistic to balance sensitivity and specificity. The approach outperformed lightweight and classical baselines in AUC, indicating superior discrimination under low illumination. These findings show that superior image quality does not always align with higher detection accuracy, and they highlight the importance of pairing effective enhancement with robust detectors for temple security. The study contributes practical insights for preserving Balinese cultural and spiritual heritage by strengthening efforts to protect pratima against theft.