<p>Distributed coordination through distributed coordination function (DCF) is a famous strategy used in a wireless local area network for shared channel access. DCF uses binary exponential backoff (BEB) to defer access to shared channels in an orderly fashion. BEB is reasonably fair. However, any organization of cooperative and collaborative agents opens up opportunities for misbehavior. A wireless agent can become greedy and start employing a strategy to win a larger share of contending trials. Traditional greedy agent detection methods lack data-driven learning. The problem domain of identification of misbehaving agents can benefit from the advent of the present data-driven and high-performance computational infrastructure. Data-driven learning can help both in modeling compelling adversaries and in counteracting them. In this paper, a greedy strategy for a rational and intentional greedy agent is devised. Following this, a greedy agent detection framework is employed to detect greedy agents using various data-driven methodologies. A new method is proposed that achieves greedy agent detection using image segmentation in a visual and intuitive manner. The proposed method is tested on two types of systems, one with a single greedy agent in a system with network churn, containing between 2 and 200 agents, and the other with 30% greedy agents in a system of 180 agents representing a highly unfair system. Formal analysis is presented for the key choices in the simulation and to better understand the results from the simulation. The proposed method is compared with state-of-the-art methods, suitable for this use case, regarding their capability to accurately classify greedy and fair agents. The comparison establishes the versatility of the proposed method in the considered systems of practical interest.</p>