The paper offers a methodology for recognizing humans by integrating the complementary features face, gait are extracted by fusion of Median Local Binary Pattern (Median-LBP) and Gabor Scale Average (GSA)-based Principal Component Analysis followed by Linear Discriminant Analysis (PCA → LDA) at decision level is proposed. Firstly, the face features and gait features are extracted using Median-LBP and GSA. In Median-LBP, the center pixel value in a 3 × 3 window is computed using every 3 × 3 neighborhood, encompassing the central pixel. In GSA, initially, a filter bank is defined using Gabor Wavelet with three scales {1, 2, 3} and six angles {0, 30, 60, 90, 120, 150}. Next, a convolution operator is used for integrating the facial and gait cues with the defined filters, and the method averages six distinct angles of Gabor matrices with similar scaling into one average imaging called GSA. Then, they are enhanced via integrating Median-LBP and GSA feature vectors denoted as Median-LBP + GSA. Subsequently, lower dimensional feature vectors are obtained with the PCA → LDA algorithm. The Euclidean distance measure is utilized for obtaining the biometric system’s face, gait decisions independently, then the AND/OR rule is employed to fuse the decisions at the decision level for recognition. The methodology is tested with the openly available face dataset (ORL) and gait dataset (CASIA B).