This paper presents a from scratch implementation and experimental analysis of the Message Digest Five hashing algorithm using C plus plus. The study examines message pre-processing, block processing, nonlinear transformation rounds, and generation of the fixed length digest. In addition to conventional text inputs, a deterministic image to text encoding pipeline is implemented using OpenCV so that image metadata and pixel values can be serialized before hashing. Experiments on modified text and image inputs show that small input changes produce substantially different digest values, demonstrating deterministic behavior and the avalanche effect. For three original and tampered image pairs, the reported digest differences correspond to Hamming distances of 65, 54, and 62 bits out of 128, respectively. The work also discusses appropriate and inappropriate uses of Message Digest Five in view of practical collision attacks. The results support its use only for non security critical integrity checks, teaching, and experimental analysis, while stronger modern hash functions should be used for security sensitive applications.

Md Afsarul Hoque, "An Image to Text Encoding Framework for Image Integrity Verification Using MD5" International Journal of Computer Science and Engineering Development, V1(4): Page(25-28) September - October 2026. ISSN: 3139-0862. www.ijcsed.com. Published by Scientific and Academic Research Publishing.