In-Cycle Optimized Sensing Point Selection-Based TFT Compensation for Enhanced Uniform Brightness OLED Displays

As the demand for display products in a wide range of sizes and formats grows, display technology has seen rapid advancements. In the case of high-end products, the ability to preserve image quality over extended periods makes it essential that brightness remains uniform throughout operation. Organic light-emitting diode (OLED) displays offer the best picture quality to date due to their self-luminous properties. However, as the usage time increases, brightness unevenness occurs due to changes in the characteristics of the driving thin film transistors (TFTs). This study proposes a measurement and signal processing system to address the degradation issue of driving-TFTs through in-cycle sensing. This method identifies the source node voltage and determines the optimal sensing point, using a driving scheme that allows sufficient time for sensing. The sensing data is used to estimate the driving-TFT characteristics with a deep learning model. In this setup, the sensing-to-inference pipeline can reach sub-millivolt accuracy and still keep the original frame time. The extra cost in performance is small. The result shows that the visual artifacts are fewer and the brightness uniformity improves, which makes the image look clearer and more stable.