This paper presents fast and robust fixed-point algorithms for independent component analysis (ICA). ICA is a statistical method for transforming observed data into statistically independent components. The authors combine two approaches: Comon's information-theoretic method and projection pursuit. They introduce a family of new contrast functions for ICA, which enable both full decomposition and individual component estimation. These contrast functions are analyzed under the linear mixture model, and methods for choosing robust or low-variance functions are discussed. The paper introduces simple fixed-point algorithms for optimizing these contrast functions, which are shown to converge quickly and reliably. The ICA problem is formulated as minimizing mutual information between transformed variables. The authors derive contrast functions based on negentropy approximations, which can be interpreted as projection pursuit directions. These functions are used to estimate independent components sequentially. The behavior of the estimators is evaluated under the linear mixture model, providing guidelines for selecting contrast functions. Practical considerations for choosing contrast functions are discussed, including statistical and numerical criteria. The paper introduces a novel family of fixed-point algorithms for optimizing contrast functions. These algorithms are shown to have fast and reliable convergence properties. Simulations confirm the effectiveness of the new contrast functions and algorithms, and real-world experiments validate their performance. The algorithms are applicable to any distribution of independent components and are robust to outliers. The fixed-point algorithms are efficient, parallel, and require minimal memory, making them suitable for real-time applications. The paper concludes that the proposed methods offer a powerful and efficient approach to ICA, with strong statistical and computational properties.This paper presents fast and robust fixed-point algorithms for independent component analysis (ICA). ICA is a statistical method for transforming observed data into statistically independent components. The authors combine two approaches: Comon's information-theoretic method and projection pursuit. They introduce a family of new contrast functions for ICA, which enable both full decomposition and individual component estimation. These contrast functions are analyzed under the linear mixture model, and methods for choosing robust or low-variance functions are discussed. The paper introduces simple fixed-point algorithms for optimizing these contrast functions, which are shown to converge quickly and reliably. The ICA problem is formulated as minimizing mutual information between transformed variables. The authors derive contrast functions based on negentropy approximations, which can be interpreted as projection pursuit directions. These functions are used to estimate independent components sequentially. The behavior of the estimators is evaluated under the linear mixture model, providing guidelines for selecting contrast functions. Practical considerations for choosing contrast functions are discussed, including statistical and numerical criteria. The paper introduces a novel family of fixed-point algorithms for optimizing contrast functions. These algorithms are shown to have fast and reliable convergence properties. Simulations confirm the effectiveness of the new contrast functions and algorithms, and real-world experiments validate their performance. The algorithms are applicable to any distribution of independent components and are robust to outliers. The fixed-point algorithms are efficient, parallel, and require minimal memory, making them suitable for real-time applications. The paper concludes that the proposed methods offer a powerful and efficient approach to ICA, with strong statistical and computational properties.