How AI Helps Michael VanPatten

Michael VanPatten is an artist who combines his passion for photography and painting with his knowledge of artificial intelligence (AI). He uses AI tools to enhance his images and create surreal effects that challenge the viewer’s perception and offer a fresh perspective on composition and color.

Michael’s art process blends the old with the new, incorporating cutting-edge AI technology with traditional techniques. He found a particular interest in exploring techniques like overlaying images to create an illusion of the subject emerging from a portal, creating a dynamic effect.

Michael’s journey as an artist began in college, where he studied art courses along with mathematics, physics, and technology. He later pursued a career in the corporate world, working as a Chief Financial Officer for several technology companies. He continued his photography work while traveling the world and capturing images of various subjects, such as beaches, boats, islands, waterfalls, marinas, golf courses, cityscapes, interesting structures, and wildlife.

After selling his most recent company in 2022, Michael returned to his artistic roots and started View Finder Arts, a digital shop art gallery named after his boat, the ViewFinder. He also acquired new cameras and equipment, such as a drone and a Go Pro, to experiment with landscape, aerial, and underwater images.

Michael’s venture into AI-driven art began two years ago, when he learned about AI’s capabilities and found interest in enhancing the surrealism of his past oil paintings with digital tools. He also used AI to breathe new life into the photographs he took decades ago, curating a collection of surreal images that reflect his vision and inspiration.

Michael currently limits his limited-edition image sales to 250 copies before retiring them. His target markets are professional office spaces, conference facilities, restaurants, hotels, galleries, institutions, and professional collectors. He also teaches part-time online business courses for Southern New Hampshire University and is a Board Member for Targa.ai, a performance Software as a Service Company.

Michael’s work illustrates the limitless potential of AI in art in a manner that reinforces traditional methods instead of decimating them. He shows how AI can help artists express their creativity and improve their artistic process, while also democratizing art and making it accessible to wider audiences through digital platform.

NST works by using a deep neural network that can extract features from images. The network consists of three components: a content image, a style image, and a generated image. The content image is the image that we want to modify, the style image is the image that we want to use as a reference, and the generated image is the image that we want to create.

The network tries to minimize two loss functions: the content loss and the style loss. The content loss measures how different the generated image is from the content image, and the style loss measures how different the generated image is from the style image. The network adjusts the pixels of the generated image until it finds a balance between the content loss and the style loss.

The content loss is computed by comparing the activations of a hidden layer in the network for the content image and the generated image. The activations represent the high-level features of the image, such as edges, shapes, and objects. The content loss is low when the generated image has similar features as the content image.

The style loss is computed by comparing the Gram matrices of multiple hidden layers in the network for the style image and the generated image. The Gram matrices capture the correlations between the activations of different filters in each layer, which represent the texture and style of the image. The style loss is low when the generated image has similar texture and style as the style image.