The search for efficient image restoration method is still a fundamental problem in image processing. Gaussian smoothing is an important linear filter smoothing method which is widely used in image denoising problem. However, outliers often remain in the edges of the recovered images obtained by linear filter smoothing. In this paper, we propose a new method which combines the Gaussian filter smoothing with L1 norm for outliers removal together for image restoration. An alternating minimization algorithm is designed to solve the proposed model. Extensive experimental results show that the proposed model is efficient for image restoration, and it greatly improves the image restoration quality in terms of both the visual effects and quantity results comparing to the Gaussian smoothing method.

The purpose is to achieve a high-quality teaching effect in quality education using the new teaching concept. Firstly, the Deep Learning (DL) theory is introduced to improve the trial-error teaching method, then the trial-error teaching is combined with STREAM education. Afterward, the conical section hyperbola teaching in college entrepreneurship education is specifically studied under experimental analysis using the proposed DL-based integrated trail-error + STREAM teaching methods. The experimental results read: student’s average veracity on the multiple-choice question is 93.4% and 90.1% for the experimental group and control group, respectively; student’s average veracity on short-answer questions is 92.3% and 90.3% for the experimental group and control group, respectively. The results show that the application of DL to the trial-error teaching method can cultivate students’ in-depth analysis and logical thinking ability for mathematical problems. Meanwhile, the DL-based integrated trial-error + STREAM teaching methods stimulate students’ initiative to learn more difficult knowledge, establish integral knowledge systems, and more comprehensively and deeply understand the teaching content. As a result, students’ scientific literacy and humanistic literacy are both improved. Therefore, the proposed DL-based integrated trial-error + STREAM teaching method in college entrepreneurship education has a guiding significance for other disciplines and provides ideas for the expansion and development of STREAM teaching in the future.

A point set registration algorithm based on improved Kullback–Leibler (KL) divergence is proposed. Each point in the point set is represented as a Gaussian distribution. The Gaussian distribution contains the position information of the candidate point and surrounding ones. In this way, the entire point set can be modeled as a Gaussian mixture model (GMM). The registration problem of two point sets is further converted as a minimization problem of the improved KL divergence between two GMMs, and the genetic algorithm is used to optimize the solution. Experimental results show that the proposed algorithm has strong robustness to noise, outliers, and missing points, which achieves better registration accuracy than some state-of-the-art methods.