Interior Design Assessment and Optimization System Using Deep Q-Feed Forward Networks

Interior Design is considered to be the most significant field that holds crucial significance for validating the comfort of the living environment. Assessing the interior design is not only a key goal but it is also crucial for visualization stages used in the design and architecture, it is also an important way to optimize the interior design plan. With the advent of Artificial Intelligence algorithms, assessment techniques have become more flexible for architects, which helps them to find innovative solutions to their various design problems. However, the existing learning algorithms suffer from the computational overhead due to their non-adaptive characteristics to the new designs, which remains be real challenge among the researchers. In this context, this research paper proposes the novel Q-learning-based Feed Forward networks for testing and validating the interior design plans with less computational overhead and higher accuracy. The comprehensive experimentation has carried out using Kaggle datasets and various performance metrics such as accuracy, precision, recall and f1-score are measured and compared with the other learning models. Results demonstrates the performance of the suggested model has achieved the accuracy of 0.97, precision of 0.965, recall of 0.964 and F1-score of 0.97. To validate the computation performance of the suggested model, model building time (MBT) was calculated in which the MBT of the suggested model has achieved 8.56secs which is better than the other existing models. Experimental outcomes illustrate the optimized assessments of interior designs systems using artificial intelligence techniques and have the potential to generate new ideas that accelerate the design process and optimizing the design scheme.