Abstract

Video traffic dominates today’s IP networks and a significant share of this traffic is generated by users accessingcontentfrom mobile devices. However, high‑quality video requires large data rates and substantial processing capability, whicharenotalways available in bandwidth‑limited and power‑constrained mobile environments. This paper presents a web‑basedarchitecturefor mobile video delivery combined with a modified decoding and compression strategy aimed at reducing bandwidthandenergyconsumption while preserving acceptable visual quality. The proposed system relies on server‑side pre‑processing, frameskippingfor low‑motion content, block‑based motion estimation, and transform coding using both the discrete cosine transform(DCT)andthe discrete wavelet transform (DWT). The web server stores both compressed and uncompressed versions of surveillance‑likevideos and streams the user‑selected version to the mobile client through HTTP. We review the theoretical foundationsoftransform‑based compression and motion‑compensated prediction, summarize key contributions fromthe literature upto2014,and describe the design of a prototype implemented in MATLAB and a web front‑end. A comparative evaluationonshortCCTV‑style clips demonstrates that the proposed DWT‑based approach achieves compression ratios above 90%withmoderatedegradation in peak signal‑to‑noise ratio (PSNR), while significantly reducing upload time and mobile data usage comparedwithDCT‑only compression. The results indicate that offloading most computational tasks to the server and tailoring compressiontolow‑motion content is an effective strategy for web‑based mobile surveillance applications. Keywords:Video compression; mobile streaming; web‑based video delivery; DCT; DWT; frame skipping; motion estimation