Camouflaged object detection (COD) has been a big challenge. It is a crucial task with numerous applications in various domains. It is contesting due to the resemblance between the object we are trying to detect and its background. Numerous methods and datasets have been introduced to tackle this problem and this domain has become one of the fastest growing domains of image processing. In this paper, we have implemented EfficientDet with SAM over a new dataset NC4K and have compared their results with some existing models. We have further analyzed the model’s failures to identify limitations and proposed potential improvements for future work. This research contributes to the exploration of EfficientDet as well as SAM by comparing it with several models for COD and the effectiveness of the new NC4K dataset.