Eos 4000d

The archaeobotanical specimens were observed and photographed using (Leica EZ4) Stereo Microscope and Canon EOS 4000D/Rebel T 100 Digital SLR Camera.
Modern dried juniper cones were collected from the best-selling herbal market in Egypt (Harraz for food Industry and Natural Products, Egypt). One of them was sectioned for comparison with the archaeobotanical specimen. The modern specimen was soaked overnight in warm water then cut with a scalpel to show the number and arrangement of seeds.
For transverse sectioning, one of the archaeobotanical specimens was embedded in a polyester resin block then cut into two halves using a jeweler saw then polished using a series of coarse- to fine-grit SiC polishing paper and ground down to 2000 grit. The same procedure was conducted for sectioning the modern juniper specimen for micromorphological examination and analysis.
The transverse section micromorphology of the archaeobotanical specimen was investigated using an environmental scanning electron microscope with energy dispersive spectroscope SEM–EDS (Quanta FEG250, with tungsten electron source, at 20 kV). Four analysis spots were analyzed. Mapping was undertaken through the scanned area to show the distribution and relative proportion (intensity) of the defined elements. The "ZAF correction method" was used for quantitative analysis of elements.

Photographs of plates were taken using a consumer-grade digital camera (Canon EOS 4000D). A total of 1,156 images of agar plates (each consisting of up to 94 strains growing in a fixed drug concentration) were generated by photographing agar dilution growth. Images were subsequently processed by splitting each plate image into 96 sub-images of 9 × 9 mm squares surrounding each spot.
These sub-images were used to create a labeled data set suitable for training an image classification algorithm. A clinical microbiologist (AG) labeled sub-images, assigning them to one of three classes based on the phenotypic appearance of colonies: (i) images with no growth (Class I), (ii) images of poor or inhibited growth (Class II), and (iii) images of uninhibited growth (Class III)—see Fig. 3 for examples. The inclusion of Class II (poor/inhibited growth) was necessary to accurately calculate MICs since colonies demonstrating antimicrobial inhibition (single colonies or a faint film) need to be disregarded (12 (link)). Image labeling took place prior to and separately from model construction. Once images were annotated, 20% of each of the three classes (no growth, poor/inhibited growth, and good growth) was reserved to test the performance of the classification models in identifying images.

The performance of the fabricated micromixers in mixing two solutions was tested using the experimental setup shown in Figure 4. First, all devices were cleaned with water and dried using warm air. Compressed air jets were used to eliminate any drops of water trapped inside the microchannels. Tape (3M Ruban Adhesif Scotch^®) was used to seal the top surfaces, thus closing the open sides of the microfluidic micromixers. Two different water streams colored yellow and blue were injected using two syringe pumps (Braintree, MA, USA, Mod. BS-300) equipped with 5-mL disposable syringes at flow rates of 50 and 100 μL/min. The streams were precisely controlled, and the flow rates of both inlets were set equally (the total flow rates in the main channel were 100 and 200 μL/min). After achieving steady-state conditions, the fluid within the microchannels was captured using a CCD camera (Canon EOS 4000D, Tokyo, Japan) with a macro lens (Canon EF 100 mm f/2.8 Macro USM, Tokyo, Japan). The captured images were then sent to a server using a USB cable for further processing by a MATLAB algorithm (see Supplementary Materials).