This paper investigates reasons for the non-utilization of ICT by SMEs in Nigeria using a survey of 180 SMEs. The results of the study indicates that majority of Nigeria SMEs utilizes basic ICT such as word processing, fixed landlines, printers and fax machines but rarely use the computer for advance functions such as business analysis, planning and decision making. However, there are key factors that inhibit these SMEs from effectively utilizing ICT in their various businesses. The survey suggests that electricity and infrastructural inadequacies are the most prevalent factors for non-utilization of ICT amongst SMEs in the region. The overall study is relevant to SMEs in general; policy makers and stakeholders in creating initiatives that will assist in the development of these SMEs.

A Facial Expression Recognition (FER) study is conducted investigating whether High Dynamic Range (HDR) tone mapped images can improve FER performance under complex lighting conditions. For this purpose, we created a new straightforward facial expression dataset of HDR images, a collection of faces under different lighting contrasts. Our approach applies a machine learning method, Support Vector Machines (SVMs) to texture-based, Local Binary Pattern (LBP) and appearance based, Speeded-Up Robust Feature (SURF) image representations to conduct emotion recognition experiments. We run comparisons using HDR tone mapped images and more traditional pre-processing methods of reducing the impact of large lighting variations. The best of the adopted methods achieved 75% and 79.8% accuracy which is significantly better than the more traditional pre-processing methods tested.

No abstract