VSvicky sharmainvickysharma.hashnode.dev·Aug 22 · 8 min readTraining the Model: Is the Recovered Data Actually Worth It?Deep in the Wild, part 7. Last of six technical posts on recovering machine-readable bird annotations from eleven years of aerial survey screenshots. Yes. A big yes, and I am very happy to share it. H00
VSvicky sharmainvickysharma.hashnode.dev·Aug 22 · 7 min readClassifying Each Dot: Giving Every Bird a SpeciesDeep in the Wild, part 6. Fifth of six technical posts on recovering machine-readable bird annotations from eleven years of aerial survey screenshots. At this point the pipeline can find the dots and 00
VSvicky sharmainvickysharma.hashnode.dev·Aug 22 · 7 min readMeasuring Honestly: Hand-Labelling 1,648 Dots to Test the DetectorDeep in the Wild, part 5. Fourth of six technical posts on recovering machine-readable bird annotations from eleven years of aerial survey screenshots. The last post ended with detection looking healt00
VSvicky sharmainvickysharma.hashnode.dev·Aug 22 · 7 min readFinding the Dots: Why Colour Detection Failed and Subtraction WorkedDeep in the Wild, part 4. Third of six technical posts on recovering machine-readable bird annotations from eleven years of aerial survey screenshots. By the end of the last post the pipeline could re00
VSvicky sharmainvickysharma.hashnode.dev·Aug 22 · 7 min readReading the Legend: Teaching the Pipeline What Each Coloured Dot MeansDeep in the Wild, part 3. Second of six technical posts on recovering machine-readable bird annotations from eleven years of aerial survey screenshots. The last post ended on one observation: everythi00
