This paper addresses the challenging task of multi-person pose estimation in real-world images, focusing on detecting and estimating the poses of multiple individuals simultaneously. The authors propose a novel approach that jointly solves the detection and pose estimation tasks by inferring the number of people, identifying occluded body parts, and disambiguating body parts between close individuals. This method contrasts with previous two-stage approaches that first detect people and then estimate their poses. The proposed formulation is an instance of an integer linear program (ILP) that implicitly performs non-maximum suppression and groups body part candidates to form consistent configurations respecting geometric and appearance constraints. Experiments on four datasets demonstrate state-of-the-art results for both single and multi-person pose estimation. The key contributions include a joint detection and pose estimation formulation cast as an ILP, and two CNN variants for generating representative sets of body part candidates. The method outperforms existing approaches in terms of accuracy and efficiency, making it a significant advancement in the field of multi-person pose estimation.This paper addresses the challenging task of multi-person pose estimation in real-world images, focusing on detecting and estimating the poses of multiple individuals simultaneously. The authors propose a novel approach that jointly solves the detection and pose estimation tasks by inferring the number of people, identifying occluded body parts, and disambiguating body parts between close individuals. This method contrasts with previous two-stage approaches that first detect people and then estimate their poses. The proposed formulation is an instance of an integer linear program (ILP) that implicitly performs non-maximum suppression and groups body part candidates to form consistent configurations respecting geometric and appearance constraints. Experiments on four datasets demonstrate state-of-the-art results for both single and multi-person pose estimation. The key contributions include a joint detection and pose estimation formulation cast as an ILP, and two CNN variants for generating representative sets of body part candidates. The method outperforms existing approaches in terms of accuracy and efficiency, making it a significant advancement in the field of multi-person pose estimation.