This paper presents survey on recent advances in machine learning (ML) and deep learning (DL) techniques for detecting and classifying lung diseases from medical imaging. Nowadays, a significant global health burden has been raised because of lung diseases, including tuberculosis, pneumonia, lung cancer, COVID-19, and so on. This article reviews methodologies employed by researchers, such as convolutional neural networks (CNNs), ensemble models, transfer learning, and data augmentation strategies, applied to chest X-rays (CXR) and CT scans. This article also reviews datasets, including NIH chest X-ray, Shenzhen, and COVID-19 CT, used to develop and validate these models, with results showing significant improvements in diagnostic accuracy. The models like DenseNet and ResNet demonstrated high AUC scores for tuberculosis, while hybrid architectures and segmentation techniques provide robust solutions for COVID-19 pneumonia detection. A transfer learning from pre-trained networks, such as AlexNet and VGG, enhanced performance especially when combined with pre-processing methods like histogram equalization and noise reduction. The review emphasizes the importance of data augmentation and hybrid model development in achieving higher accuracy and generalization for lung disease classification. This article leads to comprehensive survey of ML, DL techniques and its effectiveness in determination of lung diseases for accurate diagnosis. At the same time, it appeals the attention for future research.