In this project, a Deep Neural Network is evaluated for the task of Object Detection using low luminance dataset. Existing techniques uses a deep learning architecture to detect multiples objects during night time. In this work, a existing protocol for object detection is followed and a DNN is evaluated using Exclusively Dark Image dataset. Since the images were not annotated in the Ex-Dark dataset, a custom dataset is created using this dataset. The training images in the customized dataset are annotated using LabelImg tool. The model is trained and tested using training and testing dataset respectively. The result for this project demonstrates that the losses occuring due to low brightness are minimalized. For the evalu- ation of this model IOU metric is used and different IOU thresholds are compared to calculate the Average Precision.