Many dangerous or poisonous animals use colors or patterns to warn potential predators of danger (aposematic signal), and thus avoid molestation. Pacific newts (Taricha) produce tetrodotoxin (TTX), a paralyzing toxin, and also possess bright orange or red bellies which they reveal when confronted. However, some garter snakes (Thamnophis) have evolved resistance to this toxin. In fact, after ingesting newts, some garter snakes might retain enough TTX to be rendered poisonous themselves. We hypothesize that toxin-resistant garter snakes have evolved colors or patterns as anti-predator signals. We thus predicted that snakes with higher levels of TTX resistance would possess stronger aposematic signals (e.g., red coloration, etc). We quantified TTX resistance in snakes from across multiple sites in California (sympatric and allopatric with newts). We then quantified RGB from these same snakes from photos. We used machine learning to quantify the surface area of red coloration. We found a positive relationship between TTX resistance and red coloration. We hope to determine if this coloration is detectable by predators. This project will provide an innovative technique in machine learning to quantify a unique predator-prey system and enhance the efficiency of photograph data processing.