What is there in the 32x356 images ? are the 32x356 images just regular images of planet taken with a 2D sensor or does different pixels have anything to do with different wavelengths (Possible Approach 2 says sum up pixels along Y axis for different wavelength) ? Can anyone explain.
", "messages": [ { "id": 3014933, "postDate": "2024-10-11T18:09:47.113Z", "content": "The 32x356 arrays are not normal images; the 356-pixel dimension carries information about different wavelengths. Imagine a 1d slice of an image getting passed through a prism to separate out different colors: the 32 pixels are the spatial dimension of that 1d slice, and the 356 pixels are the different wavelengths split apart by the prism. Only a subset of 282 of those 356 wavelengths are included as part of the prediction target for this competition; see this notebook for details. Summing over the 32 pixels is one way to reduce the size and complexity of that dataset so that an analysis can focus on how things change with wavelength.
", "rawMarkdown": "The 32x356 arrays are not normal images; the 356-pixel dimension carries information about different wavelengths. Imagine a 1d slice of an image getting passed through a prism to separate out different colors: the 32 pixels are the spatial dimension of that 1d slice, and the 356 pixels are the different wavelengths split apart by the prism. Only a subset of 282 of those 356 wavelengths are included as part of the prediction target for this competition; see [this notebook](https://www.kaggle.com/code/gordonyip/update-calibrating-and-binning-astronomical-data/notebook) for details. Summing over the 32 pixels is one way to reduce the size and complexity of that dataset so that an analysis can focus on how things change with wavelength.", "votes": 2 }, { "id": 3014221, "postDate": "2024-10-11T03:51:24.227Z", "content": "What is there in the 32x356 images ? are the 32x356 images just regular images of planet taken with a 2D sensor or does different pixels have anything to do with different wavelengths (Possible Approach 2 says sum up pixels along Y axis for different wavelength) ? Can anyone explain.
", "rawMarkdown": "What is there in the 32x356 images ? are the 32x356 images just regular images of planet taken with a 2D sensor or does different pixels have anything to do with different wavelengths (Possible Approach 2 says sum up pixels along Y axis for different wavelength) ? Can anyone explain." } ], "comments": [ { "id": 3014933, "author_name": "particlebbq", "author_url": "", "post_date": "2024-10-11T18:09:47.113000", "content": "The 32x356 arrays are not normal images; the 356-pixel dimension carries information about different wavelengths. Imagine a 1d slice of an image getting passed through a prism to separate out different colors: the 32 pixels are the spatial dimension of that 1d slice, and the 356 pixels are the different wavelengths split apart by the prism. Only a subset of 282 of those 356 wavelengths are included as part of the prediction target for this competition; see this notebook for details. Summing over the 32 pixels is one way to reduce the size and complexity of that dataset so that an analysis can focus on how things change with wavelength.
", "votes": 2, "replies": [] } ], "raw_markdown_by_id": { "3014933": "The 32x356 arrays are not normal images; the 356-pixel dimension carries information about different wavelengths. Imagine a 1d slice of an image getting passed through a prism to separate out different colors: the 32 pixels are the spatial dimension of that 1d slice, and the 356 pixels are the different wavelengths split apart by the prism. Only a subset of 282 of those 356 wavelengths are included as part of the prediction target for this competition; see [this notebook](https://www.kaggle.com/code/gordonyip/update-calibrating-and-binning-astronomical-data/notebook) for details. Summing over the 32 pixels is one way to reduce the size and complexity of that dataset so that an analysis can focus on how things change with wavelength.", "3014221": "What is there in the 32x356 images ? are the 32x356 images just regular images of planet taken with a 2D sensor or does different pixels have anything to do with different wavelengths (Possible Approach 2 says sum up pixels along Y axis for different wavelength) ? Can anyone explain." } }