Figure 1.
A different view of Moore’s law.
A comparison of transistor counts in central processing units (CPUs) versus the pixel counts on cellular phone cameras. The transistor count has several data points for each year, while the cellular phone pixel count has only one data point, which corresponds to the maximum pixel count introduced in that year.
Figure 2.
A portable lensfree super-resolution colour microscope.
(a) A photograph of the microscope that weighs < 145 grams. (b) A schematic diagram of the microscope; the LED array is separated into two groups, the first group enables pixel super-resolution based on source shifting, and it contains 17 green LEDs (λ = 527 nm) where each LED is butt-coupled to a multi-mode fiber and its emission passes through a colour filter as shown in the inset. The second group of LEDs enables the acquisition of a lower resolution colour image and it is composed of three LEDs (λ = 470 nm, 527 nm and 625 nm). (c) A photograph of one of the 17 green LEDs (left) and the blue unfiltered LED (right).
Figure 3.
Image processing block diagrams.
(a) For creating a lower resolution colour image of the specimen, three lower resolution holograms are acquired, each with a different illumination wavelength (λ = 470 nm, 527 nm and 625 nm). A previous approach that simply combines these three holograms into a RGB image results in a ‘rainbow’ like colour artefact. The current approach eliminates the colour noise by averaging the colours in the YUV colour space. (b) The flowchart for acquiring and processing a super-resolved multi-height phase-recovery based grey-scale image. (c) A colour image with sub-micron spatial resolution (shown on the left) is rendered by replacing the lower resolution brightness component from (a) with the super-resolved brightness component from (b) in the YUV colour space.
Figure 4.
Quantification of the spatial resolution as a function of the number of heights used in our multi-height phase-recovery process.
(a) Amplitude image that was reconstructed using one height. (b) Amplitude image that was reconstructed using two heights. (c) Amplitude image that was reconstructed using three heights. Cross sections of the smallest resolved gratings are provided below each reconstructed image. Pixel size of the monochrome CMOS chip used in our field-portable microscope is 1.67 µm. Note that using a CMOS imager chip that has a smaller pixel size (e.g., ∼1.1 µm) [37], a higher spatial resolution of e.g., ∼300 nm can also be achieved using the same lensfree imaging technique [41].
Figure 5.
A wide FOV (∼21 mm2) lensfree colour image of a Pap smear sample (ThinPrep® preparation).
The Pap test was reconstructed using pixel super-resolved holograms acquired at four different heights (1069 µm, 1117 µm, 1159 µm and 1205 µm). For comparison purposes, 20× microscopes images (0.5 NA objective lens) of the same specimen are also provided.
Figure 6.
A wide FOV (∼21 mm2) lensfree colour image of a Pap smear sample (ThinPrep® preparation).
The Pap test was reconstructed using pixel super-resolved holograms acquired at three different heights (860 µm, 1040 µm and 1080 µm). For comparison purposes, 10× microscopes images (0.25 NA objective lens) are also provided, located above their corresponding lensfree colour images.
Figure 7.
Lensfree computational images of abnormal cells in a Pap smear sample.
(a), (b) and (c) lensfree Pap smear images of abnormal cells (pointed by white arrows), which generally exhibit large nucleus in comparison to their cytoplasm as well as an irregular cytoplasm shape. (d), (e) and (f) microscope comparison images for (a), (b) and (c), respectively, obtained using a 20× microscope objective (0.5 NA).