Surgical robot task automation has recently attracted great attention due to its potential to benefit both surgeons and patients. Reinforcement learning (RL) based approaches have demonstrated promising ability to perform automated surgical manipulations on various tasks. To address the exploration challenge, expert demonstrations can be utilized to enhance the learning efficiency via imitation learning (IL) approaches. However, the successes of such methods normally rely on both states and action labels. Unfortunately, action labels can be hard to capture or their manual annotation is prohibitively expensive owing to the requirement for expert knowledge. Emulating expert behaviour using noisy or inaccurate labels poses significant risks, including unintended surgical errors that may result in patient discomfort or, in more severe cases, tissue damage. It therefore remains an appealing and open problem to leverage expert data composed of pure states into RL. In this work, we present an actor-critic RL framework, termed AC-SSIL, to overcome this challenge of improving learning process with state-only demonstrations collected by an unknown expert policy. It adopts a self-supervised IL method, dubbed SSIL, to effectively incorporate expert states into RL paradigms by retrieving from demonstrations the nearest neighbours of the query state and utilizing the bootstrapping of actor networks. It applies similarity-based regularization and improves its prediction capacity jointly with the actor network. We showcase through experiments on an open-source surgical simulation platform that our method delivers remarkable improvements over the RL baseline and exhibits comparable performance against action based IL methods, which implies the efficacy and potential of our method for expert demonstration-guided learning scenarios. Code will be made publicly available at https://github.com/Jingshuai-cqu/AC-SSIL.

Scene text recognition (STR) methods have struggled to attain high accuracy and fast inference speed. Auto-Regressive (AR)-based models implement the recognition in a character-by-character manner, showing superiority in accuracy but with slow inference speed. Alternatively, Parallel Decoding (PD)-based models infer all characters in a single decoding pass, offering faster inference speed but generally worse accuracy. To realize the dual goals of "AR-level accuracy and PD-level speed", we propose a Context Perception Parallel Decoder (CPPD) to perceive the related context and predict the character sequence in a PD pass. CPPD devises a character counting module to infer the occurrence count of each character, and a character ordering module to deduce the content-free reading order and positions. Meanwhile, the character prediction task associates the positions with characters. They together build a comprehensive recognition context, which benefits the decoder to focus accurately on characters with the attention mechanism, thereby improving the recognition accuracy. We construct a series of CPPD models and also plug the proposed modules into existing STR decoders. Experiments on both English and Chinese benchmarks demonstrate that the CPPD models achieve highly competitive accuracy while running much faster than existing leading models. Moreover, the plugged models achieve significant accuracy improvements.

Recently, regression-based methods, which predict parameterized text shapes for text localization, have gained popularity in scene text detection. However, the existing parameterized text shape methods still have limitations in modeling arbitrary-shaped texts due to ignoring the utilization of text-specific shape information. Moreover, the time consumption of the entire pipeline has been largely overlooked, leading to a suboptimal overall inference speed. To address these issues, we first propose a novel parameterized text shape method based on low-rank approximation. Unlike other shape representation methods that employ data-irrelevant parameterization, our approach utilizes singular value decomposition and reconstructs the text shape using a few eigenvectors learned from labeled text contours. By exploring the shape correlation among different text contours, our method achieves consistency, compactness, simplicity, and robustness in shape representation. Next, we propose a dual assignment scheme for speed acceleration. It adopts a sparse assignment branch to accelerate the inference speed, and meanwhile, provides ample supervised signals for training through a dense assignment branch. Building upon these designs, we implement an accurate and efficient arbitrary-shaped text detector named LRANet. Extensive experiments are conducted on several challenging benchmarks, demonstrating the superior accuracy and efficiency of LRANet compared to state-of-the-art methods. Code is available at: https://github.com/ychensu/LRANet.git

Dominant scene text recognition models commonly contain two building blocks, a visual model for feature extraction and a sequence model for text transcription. This hybrid architecture, although accurate, is complex and less efficient. In this study, we propose a Single Visual model for Scene Text recognition within the patch-wise image tokenization framework, which dispenses with the sequential modeling entirely. The method, termed SVTR, firstly decomposes an image text into small patches named character components. Afterward, hierarchical stages are recurrently carried out by component-level mixing, merging and/or combining. Global and local mixing blocks are devised to perceive the inter-character and intra-character patterns, leading to a multi-grained character component perception. Thus, characters are recognized by a simple linear prediction. Experimental results on both English and Chinese scene text recognition tasks demonstrate the effectiveness of SVTR. SVTR-L (Large) achieves highly competitive accuracy in English and outperforms existing methods by a large margin in Chinese, while running faster. In addition, SVTR-T (Tiny) is an effective and much smaller model, which shows appealing speed at inference. The code is publicly available at https://github.com/PaddlePaddle/PaddleOCR.

Text detection in natural scenes is fundamental for text image analysis. In this paper, we propose a context-based approach for robust and fast text detection. Our main contribution is that we introduce a new concept of key region, which is described with context according to stroke properties, appearance consistency and specific spatial distribution of text line. With such context descriptors, we adopt SVM to learn a context-based classifier to find key regions in candidate regions. Therein, candidate regions are connected components generated by local binarization algorithm in the areas, which are detected by an offline learned text patch detector. Experimental results on two benchmark datasets demonstrate that our approach has achieved competitive performances compared with the state-of-the-art algorithms including the stroke width transform (SWT) [1] and the hybrid approach based on CRFs [2] with speedup rates of about 1.7x~4.4x.

In this paper, we propose a method for dot text detection based on FAST points. This problem is different from general scene text detection because of discontinuous text stroke. Unlike many other methods which assume that text is horizontally oriented, our method is able to deal with slant dot text. We extract interesting patches from FAST points and define four features based on the stroke and gray value similarity of dot text to describe a patch. Then, we generate some candidate regions from these patches and utilize SVM to filter out non-dot text ones with the first and second order moments of FAST points in them. Experimental results show that the proposed method is effective and fast to detect dot text.