Utilizing a digital canvas to conduct a spatial-semantic search for digital visual media

In a digital medium environment, a computer-implemented method of searching for and identifying digital images based on semantic and spatial information, comprising:

• receiving user input of a query term and query area corresponding to a digital canvas, wherein the query term indicates targeted visual content and the query area indicates a targeted region for portraying the targeted visual content;
• determining a query feature set based on the query term and the query area, wherein determining the query feature set comprises:
• generating a representation of the query area and the query term;
• and
• providing the representation of the query area and the query term to a query neural network;
• and
• identifying, from a plurality of digital images, a digital image portraying the targeted visual content within the targeted region by comparing the query feature set to feature sets learned from the plurality of digital images using a digital image neural network.

A system for identifying digital images based on semantic and spatial information, comprising:

• one or more memories comprising:
• a plurality of features sets, wherein each feature set:
• corresponds to a digital image of a plurality of digital images, and
• is extracted from a layer of a digital image neural network that preserves semantic and spatial information from the corresponding digital image;
• a query neural network trained to generate query feature sets from representations of query areas and query terms, the query feature sets having a dimensionality of the features sets of the plurality of digital images;
• at least one server storing instructions thereon, that, when executed by the at least one server, cause the system to:
• generate a representation of a query area and a query term that encodes the query term at a spatial location corresponding to the query area, wherein the query term indicates targeted visual content and the query are indicated a targeted region for portraying the targeted visual content;
• generate, using the query neural network, a query feature set from the representation of the query area and the query term;
• and
• identify, from the plurality of digital images, a digital image portraying the targeted visual content within the targeted region by comparing the query feature set with the plurality of feature sets.