There is a growing thrust of studies on exploring applications of multiwavelets in image compression, denoising and recognition. For compression, multiple scaling functions and multiple wavelets provide more flexible approximation of the original signal at high compression ratio than a comparable single scaling function and single wavelet, leading to potential improvement in the detection of some features even in low bit-rate videos. This paper presents a study on detection of human skin colors in low bit-rate videos as a performance evaluation of low bit-rate videos encoded using spatio-temporal 3-D multiwavelets.