As real-time object detection systems such as autonomous cars need to process input images from multiple cameras and perform ML-based inferencing in real-time, they face significant challenges in delivering accurate and timely inferences. To meet this need, we aim to provide different levels of detection accuracy and timeliness to different portions with different criticality levels within each input image. We develop DNN-SAM, a dynamic Split-And-Merged DNN (deep neural network) execution and scheduling framework that enables seamless Split-And-Merged DNN execution for unmodified DNN models. Instead of processing an entire input image once in a full DNN model, DNN-SAM first splits a DNN inference task into two smaller sub-tasks—a mandatory sub-task dedicated to a safety-critical (cropped) portion of each image and an optional sub-task for processing a down-scaled image—then executes them independently, and finally merges their results into a complete one. To achieve timely and accurate detection for DNN-SAM, we also develop two scheduling algorithms that prioritize sub-tasks according to their criticality levels and adaptively adjust the scale of the input image to meet the timing constraints while minimizing the response time of mandatory sub-tasks or maximizing the accuracy of optional sub-tasks. We have implemented and evaluated DNN-SAM on a representative ML framework. Our evaluation shows DNN-SAM to improve detection accuracy in the safety-critical region by 2.0–3.7× and lower average inference latency by 4.8–9.7× over existing approaches without violating any timing constraints.